<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Solitude Agents - Medium]]></title>
        <description><![CDATA[A blog series following all things related to LLM Agents, Generative AI and and workflow automation like RPA and Process Mining. - Medium]]></description>
        <link>https://medium.com/solitude-agents?source=rss----a1447ff6a896---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Solitude Agents - Medium</title>
            <link>https://medium.com/solitude-agents?source=rss----a1447ff6a896---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 19:40:40 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/solitude-agents" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Fraud is destroying small 3PLs — TIA]]></title>
            <link>https://medium.com/solitude-agents/fraud-is-destroying-small-3pls-tia-b869d18687d4?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/b869d18687d4</guid>
            <category><![CDATA[freight-forwarding]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[logistics]]></category>
            <category><![CDATA[3pl]]></category>
            <category><![CDATA[fraud-detection]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Tue, 12 Aug 2025 15:26:43 GMT</pubDate>
            <atom:updated>2025-08-12T12:02:56.460Z</atom:updated>
            <content:encoded><![CDATA[<h3>Fraud is destroying small 3PLs — TIA</h3><figure><img alt="An accounts payable executive stressed about the volume of documents." src="https://cdn-images-1.medium.com/max/1024/1*VDzNukzeRc7l2ODxvu2_WA.png" /><figcaption>A person overwhelmed with invoice paperwork.</figcaption></figure><p>The TIA is sending a clear message: criminals posing as legitimate carriers are sending fake invoices and POs to honest brokers and killing their businesses. Smart business owners are utilizing AI to prevent $ 200,000 in annual losses due to fraudulent invoices.</p><p>Any savvy business owner knows that they live and die by their cash flow statements. When it comes to collecting your money, you don’t play around! That’s why processing carrier invoices and PODs as quickly as possible is critical to hitting your numbers on time at the end of the month.</p><p>So, where can this go wrong? When business is booming and we have 10, 20, even 30 carrier invoices flying in, it gets harder and harder to check every detail while staying on top of your day-to-day business.</p><p>Sometimes an ‘o’ can look like a zero, PO numbers don’t match, or the signature from your carrier is missing on that Bill of Lading. In a hurry, you might file the invoice away quickly so you can move on and get paid by the customer as soon as you can.</p><p>Bad news is, scammers know this, and they’re taking advantage of you. A TIA survey reported that <strong>22% of respondents lost $200k or more</strong> to ‘strategic fraud’. <strong>Truckload freight accounts for 97% of all cases.</strong></p><figure><img alt="An infographic illustrating fraud issues for freight brokers." src="https://cdn-images-1.medium.com/max/1024/1*JH613isDnANT1zmWxJS1QQ.png" /><figcaption>A TIA infographic illustrating the effect of fraud on freight brokers in the USA.</figcaption></figure><p>These fraudsters have a simple playbook. Send legitimate-looking invoices and PODs to small brokers who are overwhelmed with paperwork, hoping someone makes a mistake and approves the payment.</p><p>Businesses are trying to combat this, spending $199k or more on fraud prevention tools. The only issue is this: all that extra spending eliminates those already thin margins you’ve been fighting to grow.</p><p>Here’s the good news: <strong>new tools are getting the job done at only 95% the cost of traditional tools.</strong> It’s everyone’s favourite buzzword lately, AI. Not just any AI, though; these new tools are built specifically for freight forwarders and 3PL providers.</p><p>Take it from our customer: Legion Logistics. Using our AI, they’ve been able to check every invoice as it comes into their email and approve it into their TMS with the click of a button. No more manual checks, no more second-guessing.</p><p>If the PO isn’t in your system or something doesn’t match, our AI instantly flags it and lets you decide what to do. Being in the driver’s seat of your business is what got you here, so it shouldn’t be different when it comes to your accounts.</p><p>Want to learn more? Book a call to see a demo here: <a href="https://www.solitude.ai/book-a-demo">https://www.solitude.ai/book-a-demo</a>.</p><h4>Sources:</h4><ol><li><a href="https://www.dcvelocity.com/editorial/featured/freight-brokers-battle-rising-fraud-schemes">https://www.dcvelocity.com/editorial/featured/freight-brokers-battle-rising-fraud-schemes</a></li><li><a href="https://landline.media/fmcsa-sharpening-focus-on-broker-fraud/">https://landline.media/fmcsa-sharpening-focus-on-broker-fraud/</a></li><li><a href="https://www.overdriveonline.com/business/article/15684022/brokers-under-seige-at-pivotal-moment-with-freight-fraud">https://www.overdriveonline.com/business/article/15684022/brokers-under-seige-at-pivotal-moment-with-freight-fraud</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b869d18687d4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/fraud-is-destroying-small-3pls-tia-b869d18687d4">Fraud is destroying small 3PLs — TIA</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Should your business be using AI agents?]]></title>
            <link>https://medium.com/solitude-agents/should-your-business-be-using-ai-agents-086b5c5881ed?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/086b5c5881ed</guid>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[automation]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Thu, 06 Jun 2024 00:25:19 GMT</pubDate>
            <atom:updated>2024-06-06T00:40:45.864Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*st8sptwx2dx6_c4u.jpg" /><figcaption>Source: <a href="https://blog.upal.com.my/why-hard-work-is-not-enough-to-succeed-at-work/">https://blog.upal.com.my/why-hard-work-is-not-enough-to-succeed-at-work/</a></figcaption></figure><h3>Introduction</h3><p>You’re probably spending over 20 hours a week carrying out monotonous tasks like responding to emails and entering data into various systems. We all wish we could spend that time on important strategic work that could grow our businesses without being burdened with tedious labor that clogs up our daily calendars.</p><p>Through our interviews with $20–100 million ARR businesses, we reached the same conclusion as a 2023 study by <a href="https://www.techradar.com/pro/workers-are-wasting-hours-every-day-on-menial-it-tasks#:~:text=Workers%20are%20spending%20as%20much,Benchmark%20Report%20by%20Kore.ai.">techradar.com</a> which found that we spend “as much as half [our] day (3–4 hours) doing non-productive work or manual tasks that could be automated”. If drafting and responding to emails or manually entering data into a spreadsheet / CRM / ERP wasn’t your childhood dream, you’re certainly not alone. Even though they are crucial, these tasks steal your team’s peak productivity hours while preventing them from focusing on strategic, high-value work. Our customer discovery sessions revealed a common pain point:</p><p>It is extremely difficult to automate high-impact tasks which require significant human input. This is because they rely on unstructured data like PDFs, Excel Sheets, handwritten documents, images, or contracts.</p><p>For the first time, technology has enabled a one-size-fits-all solution that can work for your business off-the-shelf: AI Agents.</p><p>We break down AI agents in detail <a href="https://medium.com/solitude-agents/what-is-an-llm-agent-for-non-techical-people-2baec0326583">here</a> for those interested, otherwise here’s a brief description: AI agents are programs that automate tasks that involve data interpretation, tool use, and human approval by leveraging large language models (LLMs), machine learning algorithms, and guidance frameworks to complete domain-specific workflows. These agents can be integrated into your daily routine simply by sending them an email in plain language assigning the work.</p><h3>Where can you use AI agents in your business?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*c_UTxeObpmWYW9nPMEFzcQ.png" /><figcaption>The Solitude work scheduler for Agents.</figcaption></figure><p>Let’s dive into some real-world examples where AI agents can significantly impact your business operations. Consider tasks like bank reconciliations or invoicing. While essential for maintaining financial accuracy, they can be time-consuming and costly. As mentioned, they can take up to 20–30 hours per week of your team’s time or even require a dedicated full-time employee. By automating these processes with AI agents, you save hundreds of hours annually and can redirect your employees’ efforts towards more strategic and value-adding activities, such as improving customer relationships, optimizing operational processes, or developing new products and services.</p><p>With an AI agent, you simply connect it to your email and ERP software, assign the task, and let the agent do the rest. The agent will read the invoices, extract the relevant information, and input it into your ERP system — all without any manual intervention.</p><p>AI agents help streamline critical tasks that span multiple areas of your business and demand substantial due diligence to prevent errors. This can include preparing financial reports for investors or analyzing large datasets to identify market trends. Each of these requires significant time investments from senior executives, easily consuming over 40–50 hours per month. This diverts their focus away from strategic decision-making and leadership responsibilities such as setting long-term goals, fostering a positive company culture, or mentoring and developing team members.</p><p>By leveraging AI agents through a simple task assignment interface, you can automate data collection, analysis, and reporting. Agents enable team members to make informed decisions faster than ever before. This not only reduces the risk of human error but also allows your leadership to allocate their time more effectively, focusing on initiatives that directly contribute to the success and growth of your organization.</p><h4>Before you carry on…</h4><p>If you’ve enjoyed reading to this point, please consider signing up to our waitlist to stay up to date for more content like this:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fdocs.google.com%2Fforms%2Fd%2Fe%2F1FAIpQLSc7rARs1OWgWRu-Sf8eZR3j7Bj5YmsZxfe7WT8J7cF7xyYLtA%2Fviewform%3Fembedded%3Dtrue&amp;display_name=Google+Docs&amp;url=https%3A%2F%2Fdocs.google.com%2Fforms%2Fd%2Fe%2F1FAIpQLSc7rARs1OWgWRu-Sf8eZR3j7Bj5YmsZxfe7WT8J7cF7xyYLtA%2Fviewform%3Fusp%3Dsend_form&amp;image=https%3A%2F%2Flh6.googleusercontent.com%2FQ2PerAes2LNlfT5coQop0tXRVbKOf6NczyiRPpDFmPalW48MaKNJDxoa5Rbn-nKnX4PtoP7xXP4%3Dw1200-h630-p&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=google" width="760" height="675" frameborder="0" scrolling="no"><a href="https://medium.com/media/147c86f4bd4b71ce4e976fa658236f42/href">https://medium.com/media/147c86f4bd4b71ce4e976fa658236f42/href</a></iframe><h3>What is the hidden cost of manual repetitive tasks?</h3><p>It’s not just about the time and resources consumed by these manual tasks — it’s also about the opportunity cost. Every hour not spent on devising and executing strategic plans is lost revenue. This was demonstrated in a study by <a href="https://spaexecutive.com/2023/10/06/the-business-cost-of-manual-processes-may-be-25-of-your-revenue/">spaexecutive.com</a> which showed that <strong>manual tasks could be costing your business up to 25% of your revenue.</strong> By automating these tasks with AI agents, you’re unlocking your team’s potential and positioning your business for growth.</p><p>When implementing AI agents, it’s essential to consider data security and the quality of work produced. A reliable AI platform should prioritize secure connections between AI agents and your internal tools, ensuring that your sensitive data remains protected, especially if you work in regulated environments like Finance and Healthcare. LLM API providers must also facilitate your data locality requirements, ensuring compliance with relevant regulations and standards. The platform should implement proper data storage and processing protocols to maintain data integrity and confidentiality. As a business, you must confirm these requirements are met before you trust an application with your sensitive data.</p><p>Businesses that perform repetitive, manual tasks involving unstructured data and requiring a lot of human effort will benefit most from AI agents. As your business grows, communication and alignment become increasingly complex, making it challenging to maintain cross-team collaboration and consistency across work output. AI agents can mitigate those pain points across the board by being able to access the necessary data constantly and scaling instantly according to the tempo of your business. This ensures that your processes will maintain the same level of output as your operations expand and evolve, allowing your teams to focus on their most important objectives such as improving customer satisfaction, optimizing supply chain management, or developing new products or services tailored to your target market.</p><h3>Is now the right time to adopt AI Agents?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/941/0*4vlXVQpQK6xcyK6Q" /><figcaption>Source: <a href="https://www.business-planning-for-managers.com/main-courses/marketing-sales/marketing/the-adoption-curve/">https://www.business-planning-for-managers.com/main-courses/marketing-sales/marketing/the-adoption-curve/</a></figcaption></figure><p>The adoption of AI agents is rapidly accelerating, with businesses across industries recognizing their transformative potential. According to a recent survey by McKinsey, <strong>50% of companies have already adopted AI in at least one business function, and this number is expected to grow to 80% by 2025.</strong> As the technology matures and becomes more accessible, the barriers to entry are falling away, making it easier and more cost-effective for businesses to implement customizable generative AI solutions.</p><p>One of the key drivers behind this increased accessibility is the reduction in costs associated with developing and deploying AI agents. As more players enter the market and competition increases, businesses can now access powerful AI capabilities at a fraction of the cost compared to just a few months ago. The prices of these services will only trend downward as LLMs become increasingly commoditized.</p><p>Today, even smaller companies with limited budgets can now harness the power of AI to automate tasks, create easily customizable workflows that suit the firm’s needs, and gain a competitive edge.</p><p>Moreover, the level of customization and flexibility offered by modern AI platforms allows businesses to tailor AI agents to their specific use cases and workflows. Whether you need an AI agent to assist with customer support, data analysis, or content creation, you can now configure and deploy these agents to seamlessly integrate with your existing tools and processes. This level of adaptability ensures that AI agents can deliver maximum value to your business without requiring significant changes to your current operations.</p><p>As the barriers to entry continue to fall and more businesses adopt generative AI solutions, those who fail to keep pace risk being left behind. By embracing AI agents now, you can position your company at the front of this technological revolution and become one of the first adopters to reap the benefits of this technology, a technological shift that is easily on par with the mass adoption of the internet itself in the early days.</p><h3>Conclusion</h3><p>In conclusion, the rise of AI agents represents a pivotal moment in the evolution of business technology. As we’ve seen, the challenges faced by enterprises across industries — from time-consuming manual tasks to the increasing complexity of operations — can be effectively addressed by the strategic implementation of AI agents. The barriers to entry are lower than ever, and the potential benefits are vast.</p><p>However, it’s crucial to recognize that this window of opportunity won’t remain open indefinitely. As more and more businesses adopt AI agents and integrate them into their workflows, those who hesitate risk falling behind in an increasingly competitive landscape. The time to act is now.</p><p>By understanding the current capabilities and future potential of AI agents, you can make informed decisions about how to leverage this technology to solve your specific pain points and position your business for long-term success. Whether you’re looking to automate repetitive tasks, streamline complex processes, or gain deeper insights from your data, AI agents offer a powerful solution.</p><p>So, if you’re ready to take your business to the next level and stay ahead of the curve in the AI-driven future, now is the time to explore how AI agents can transform your operations. Don’t let this opportunity pass you by — the future of your business depends on it. As always, if you’ve enjoyed reading this, signing up to the waitlist on our website is the best way to stay up to date for content like this: <a href="https://solitude.ai/">https://solitude.ai</a></p><h4>Sources</h4><ol><li><a href="https://www.techradar.com/pro/workers-are-wasting-hours-every-day-on-menial-it-tasks">https://www.techradar.com/pro/workers-are-wasting-hours-every-day-on-menial-it-tasks</a></li><li><a href="https://spaexecutive.com/2023/10/06/the-business-cost-of-manual-processes-may-be-25-of-your-revenue/">https://spaexecutive.com/2023/10/06/the-business-cost-of-manual-processes-may-be-25-of-your-revenue/</a></li><li><a href="https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2023-generative-AIs-breakout-year">https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2023-generative-AIs-breakout-year</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=086b5c5881ed" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/should-your-business-be-using-ai-agents-086b5c5881ed">Should your business be using AI agents?</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Startups still haven’t cracked distribution for AI Agents]]></title>
            <link>https://medium.com/solitude-agents/startups-still-havent-cracked-distribution-for-ai-agents-6023ee732234?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/6023ee732234</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[technology-adoption]]></category>
            <category><![CDATA[distribution]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[ai-agent]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Thu, 16 May 2024 00:55:40 GMT</pubDate>
            <atom:updated>2024-05-19T20:17:45.915Z</atom:updated>
            <content:encoded><![CDATA[<h3>What is the missing ingredient for AI Agent adoption?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*G1ABceIz92FBcxko.jpg" /><figcaption>Distribution is central to any business or product, AI Agents are no different! Source: <a href="https://www.revolutiontrucking.com/blog-posts/freight-truck-different-types-and-what-they-haul">https://www.revolutiontrucking.com/blog-posts/freight-truck-different-types-and-what-they-haul</a></figcaption></figure><h3>LLM Agents are still very early in the technology adoption curve</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*sjPQmpVFh420J1-V" /><figcaption>Source: <a href="https://ankitmaurya-uxfoot.medium.com/technology-adoption-curve-c2b7f108e922">https://ankitmaurya-uxfoot.medium.com/technology-adoption-curve-c2b7f108e922</a></figcaption></figure><p>Businesses are looking to onboard generative AI to solve specific pain points but they often don’t know where to start or which use cases to prioritize (more on that <a href="https://medium.com/solitude-agents/how-to-build-a-llm-agent-for-enterprises-2ba054dfa17c">here</a>). In light of the recent GPT-4o announcements, agents and agentic experiences are becoming increasingly popularised in the public sphere. Even so, businesses looking to onboard generative AI into their business just can’t seem to find the right match for their needs, what gives?</p><p>Although there have been some large leaps in the progression of agent technology and foundational LLMs, there’s still a wide gap in the space of applications that are relevant to businesses today.</p><blockquote>A study by Salesforce showed that 62% of workers felt like they lacked the skills to “use [generative AI] technologies accurately and safely”.</blockquote><p>This means there is still a significant user experience gap for agent applications to overcome before they can be considered a mainstream application. To us, it seems clear that the shortfall is more of a user experience (UX) challenge, rather than a technological one given that AI startups have raised a collective $18 billion this year alone; mostly on good traction in the B2B sector.</p><p>It seems like AI is going through growing pains as it slowly begins to work its way into the early adopter band of the technology adoption curve. However, to cross the famous gap that exists between early adopters and the early majority, as a community we have to address the user experience challenges that are plaguing the current suite of tools. Using Marc Andreessen’s analogy of AI as a type of computer, at the moment it still feels like customers have to interface with these tools using punch cards, which is unthinkable when compared to the seamless experience we take for granted with our laptops today.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/631/0*3Gof-FG1JndtD8SN" /><figcaption>On the left, is a CDC-6500. On the right, is a device with more than 100x computing power compared to the CDC. Source L: <a href="https://blog.wirelessmoves.com/2019/01/chasing-seymour-supercomputing-and-punchcards.html">https://blog.wirelessmoves.com/2019/01/chasing-seymour-supercomputing-and-punchcards.html</a>, Source R: <a href="https://www.trustedreviews.com/reviews/macbook-air-m1">https://www.trustedreviews.com/reviews/macbook-air-m1</a></figcaption></figure><h3>Distribution for AI Agents is a big problem for builders</h3><p>Even though the UX for many of today’s AI Agent applications leaves a lot to be desired, it presents a massive opportunity for innovation and ideation for agent builders to discover. While builders can focus on developing agents that can attend to a specific need, they are faced with another problem: Once you build your agent, how do you get it into the hands of your customers?</p><p>On the surface the solution seems obvious; build a SaaS and market that to your target customer profile. Unlike a traditional software business, however, some gaps need to be filled in first.</p><h4>Do your customers know what they’re looking for?</h4><p>The businesses that we’ve spoken to have all indicated that they understand the potential generative AI offers their business, but they’re not quite sure where or how to leverage it. This indicates a customer that is problem-unaware. This means despite your best SEO efforts, it’s unlikely that the customer will find your solution by direct search online.</p><p>If this rings a bell then you aren’t alone; the agent builders we’ve spoken to all listed distribution and customer acquisition as their main problems. Often data locality is an afterthought as they are very much focused on getting the business off the ground for the first agent use case that they’ve built out.</p><h4>Can your LLM agent integrate with the customer’s existing software?</h4><p>To provide the most value to the customer, your agent needs to be able to integrate nicely with their current software ecosystem. Every business uses its software stack (SAP vs Microsoft Dynamics for example). Even if your agent is generally applicable to a particular task, you might have to change your prompt or add new tool integrations to support different kinds of customers. Given the software diversity amongst businesses, there’s a high chance you have to do this often to support all the different permutations of the stack used to complete the same process (think of all the ways and different software you can use to reconcile invoices).</p><h4>Can your agent work in the customer’s locality?</h4><p>If the customer is in the EU or UK, then they will likely require that all their data is processed and stored in their data locality. Depending on your tech stack, this can be a tall order! Especially since many LLM API providers do not explicitly make any data residency guarantees, especially around data processing.</p><h3>Businesses want generative AI but don’t know where to use it</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*tpQ_NX9th8ywyy5o" /><figcaption>Source: <a href="https://www.gong.io/fr/blog/sales-funnel-examples/">https://www.gong.io/fr/blog/sales-funnel-examples/</a></figcaption></figure><p>B2B buyers rely mostly on reviews to make purchases, but right now, there isn’t anywhere they can go to compare AI solutions for their business.</p><blockquote>A recent Gartner study shows that between 37–41% of buyers rely solely on reviews before buying new software products, and these “review-reliant” buyers are 70% more likely to make a software update as compared to their counterparts.</blockquote><p>It’s important to note that agents present an upgrade to the way the customer currently accomplishes a task, and the main blocker of automation is simply resistance to change. Arming your potential customer’s champion (the warm lead on your site) with all the data they need to make a purchase decision is critical.</p><p>If you’ve managed to generate interest in your product and attract qualified leads who understand what they’re buying, the next step of the process is<strong> evaluation</strong>. As part of the B2B buying process, the customer’s next step will be to compare your solution against others on the market. It is important in the B2B buying process to be able to conduct a full and thorough evaluation of a purchase and its alternatives. As a business customer, they must carefully consider the reputational risk should a software purchase not pan out. More generally, we are more careful in the workplace regarding purchases that might affect the business because our credibility with our peers is at stake; this is one of the key differentiators between consumer and business buying habits.</p><p>The big question is, where can your customers go to compare and evaluate different agents / generative AI solutions for their use case? At the moment, short of reviewing the highly technical metrics from academic benchmarks like AgentBench and ChatArena, there is no straightforward way for the customer to make an apples-to-apples comparison of different agentic solutions to determine the fit for their business. As of right now, most business customers would have to either hire or sub-contract the talent that would be able to make the distinction between different agent products. Just think, if you were to google (or search on perplexity) for an agent that could summarise emails, book appointments, and enter data in a CRM, how would you compare agents from different vendors?</p><p>The AI sales agent I just described is a popular use case and is already on offer by multiple startups (regie.ai, taskade.com, and b2brocket.ai to name a few). All have different pricing and overlapping features. They likely utilize different LLMs which might be a mix of closed-source, open-source, and fine-tuned variants. As a customer, testing all these different solutions could take months, and I might fall out of the funnel entirely as my organization is already resistant to change.</p><h3>How to distribute LLM-based AI Agents</h3><p>Our research shows that businesses are not searching for “AI agents”, nor should they!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*glNlIKuNs8aDCfmIDX-lIw.png" /><figcaption>Search trends for AI Agent. Source: <a href="https://trends.google.com/trends/explore?date=today%205-y&amp;q=AI%20Agent">https://trends.google.com/trends/explore?date=today%205-y&amp;q=AI%20Agent</a></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*33f6KYa8kvRmHIEWGtTigg.png" /><figcaption>Search trends for Automation (in red) compared to AI Agent (in blue). Source: <a href="https://trends.google.com/trends/explore?date=today%205-y&amp;q=AI%20Agent,Automation">https://trends.google.com/trends/explore?date=today%205-y&amp;q=AI%20Agent,Automation</a></figcaption></figure><p>As builders ourselves, we’ve been trying many different methods to understand what works and what doesn’t for agent distribution and distilled them into the points below:</p><ol><li><strong>Market your agent as an automation solution for a specific, niche problem within a business.</strong> Businesses have pains that they want to solve today, and generative AI is one of many possible solutions; they’re more likely to be searching for questions like “How do I automate email scraping?” or “Automatically reconcile invoices in SAP”. Businesses are struggling with manual, repetitive tasks that take up more than 20 hours per week. These tasks have been notoriously difficult to automate because most of them involve unstructured data like emails and PDFs, a perfect fit for an agent! Positioning your agent as a solution to specific problems significantly increases your chance of your solution getting found and resonating with your target customer.</li><li><strong>Onboard agents into a customer’s ecosystem is by providing native software integrations with their current technology stack.</strong> As part of your agent’s toolkit, it should be able to reach out and interact with the customer’s daily drivers like Excel, Quickbooks, SAP, Outlook, and many more. By integrating with their existing tools, you can piggyback off the adoption of these more widely adopted platforms to help onboard your agent into these businesses.</li><li><strong>If you’re using a closed-source LLM, research their data policies closely to understand how and where they process data.</strong> Businesses, in particular enterprises, will be very sensitive to this and will ask for guarantees on these policies. For instance, together.ai and OpenAI mostly process and store all data in the USA, which is fine for US-based customers but not for the EU or the UK. On the other hand, VertexAI (which supports the Google Gemini models), Azure OpenAI, and Mistral.ai provide guarantees for EU data residency and GDPR, which will let you serve customers in these localities. For those building on self-hosted, open-source LLMs, you have the much more difficult engineering challenge of ensuring that your model is deployed into the right regions and the data is processed right next to it. This can create complexities around routing, serving, and storing data from an infrastructure perspective.</li></ol><h3>How Solitude manages distribution and infrastructure for LLM-based AI Agents</h3><p>The guidance above is easier said than done. On the Solitude platform, we wanted to make the lives of agent builders way easier by handling all the above for you. For agents built and posted on the Solitude marketplace, we’re developing the following feature-set:</p><ol><li><strong>Managed infrastructure.</strong> Unique to Solitude, we manage all the agent infrastructure on behalf of both you and your customers to facilitate a smooth onboarding experience. Solitude hosts the agents on behalf of businesses either on our SaaS platform or in Bring-Your-Own-Cloud style deployments if the business has an on-prem requirement.</li><li><strong>Lead generation</strong>. We also carry out direct sales outreach to attract businesses with $10m+ in revenue who have specific pains around automation and manual data entry work and connect them directly with your agent. We will SEO optimize all the agent postings on the marketplace, and then cross-post them on sites like Upwork &amp; Fiverr in response to job listings for workflows you indicate the agent could fulfill.</li><li><strong>Managed integrations.</strong> Solitude provides pre-made software integrations with common business applications like SAP, Outlook, Microsoft Office, Salesforce, Hubspot, and much more. We let you develop agents to automate specific workflows using these tools and test different planning algorithms like Tree-of-thought, React, and Graph of Thought to compare which ones work the best for different workflows.</li><li><strong>Seamless data residency.</strong> While you can select a preferred LLM for a particular agent/workflow, we also let you select alternatives in case LLM is not available in the customer’s data locality. We also manage all files and data uploaded by customers to our platform to ensure that they’re stored in a way that is compliant with the customer’s local data regulations.</li></ol><h3>Summary</h3><p>Have you built an agent for a problem that you or your team faced and think that others would find it useful? We think you’re right! We’ve come across many amazing builders in the space who are creating immensely useful applications, but don’t have a way of getting them into the hands of customers. We built the Solitude platform for you.</p><p>If you’ve found this insightful and want to get in early on a platform that lets you get paid for your agents, signing up for our waitlist below or on our website is a great place to get an invite: <a href="https://solitude.ai/">https://solitude.ai</a></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fdocs.google.com%2Fforms%2Fd%2Fe%2F1FAIpQLSc7rARs1OWgWRu-Sf8eZR3j7Bj5YmsZxfe7WT8J7cF7xyYLtA%2Fviewform%3Fembedded%3Dtrue&amp;display_name=Google+Docs&amp;url=https%3A%2F%2Fdocs.google.com%2Fforms%2Fd%2Fe%2F1FAIpQLSc7rARs1OWgWRu-Sf8eZR3j7Bj5YmsZxfe7WT8J7cF7xyYLtA%2Fviewform%3Fusp%3Dsend_form&amp;image=https%3A%2F%2Flh6.googleusercontent.com%2FQ2PerAes2LNlfT5coQop0tXRVbKOf6NczyiRPpDFmPalW48MaKNJDxoa5Rbn-nKnX4PtoP7xXP4%3Dw1200-h630-p&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=google" width="760" height="675" frameborder="0" scrolling="no"><a href="https://medium.com/media/6e609a6102b45d79bf3914c78bc4c0f4/href">https://medium.com/media/6e609a6102b45d79bf3914c78bc4c0f4/href</a></iframe><h3>Sources</h3><ol><li><a href="https://www.gartner.com/en/digital-markets/insights/review-reliant-buyers">https://www.gartner.com/en/digital-markets/insights/review-reliant-buyers</a></li><li><a href="https://www.signavio.com/post/change-at-work/">https://www.signavio.com/post/change-at-work/</a></li><li><a href="https://www.salesforce.com/uk/news/stories/generative-ai-ethics-survey/">https://www.salesforce.com/uk/news/stories/generative-ai-ethics-survey/</a></li><li><a href="https://www.chiefaioffice.xyz/c/database?utm_source=www.chiefaioffice.xyz">https://www.chiefaioffice.xyz/c/database</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6023ee732234" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/startups-still-havent-cracked-distribution-for-ai-agents-6023ee732234">Startups still haven’t cracked distribution for AI Agents</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to build a LLM agent for Enterprises]]></title>
            <link>https://medium.com/solitude-agents/how-to-build-a-llm-agent-for-enterprises-2ba054dfa17c?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/2ba054dfa17c</guid>
            <category><![CDATA[automation]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[llm-agent]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Thu, 02 May 2024 19:01:49 GMT</pubDate>
            <atom:updated>2024-05-05T13:38:17.262Z</atom:updated>
            <content:encoded><![CDATA[<h3>What problems should LLM Agents be used for?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*9OtlRH7nfbSCAr7T" /><figcaption>Successful agent implementations for enterprise requires the delicate orchestration of distributed systems in order work as intended. It’s hard to get right! Source: <a href="https://asuevents.asu.edu/event/asu-symphony-orchestra-picturesque">https://asuevents.asu.edu/event/asu-symphony-orchestra-picturesque</a></figcaption></figure><p>We conducted customer discovery sessions with the operators of multi-million dollar revenue businesses across verticals like media, finance, and utilities; only to discover that all their productivity problems clustered around a core theme. All these businesses experienced severe pain when they attempted and failed to automate high-impact tasks that require significant manual effort with timescales ranging from hours to days to complete. More often than not, these problems were a perfect match for agentic applications as they required both data interpretation, tool use, and human approval in some cases.</p><p>With LLMs becoming cheaper, more complex automation use cases become viable, including the rapid emergence of agent architectures for the most advanced tasks. Despite advancements in technology, it is still unclear how businesses are expected to consume and implement agents that work on their private data or require access to internal tools.</p><h3>Finding the right problems for LLM Agents to solve</h3><p>Before diving deeper, it is important to acknowledge the spectrum of tasks that are available for automation within a business, as well the relative value that automation may hold:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ymugQ8pvSU_apTO-N5cdoQ.png" /><figcaption>A graph of impact vs cost for automations within an enterprise. Inspired by the amazing work of <a href="https://twitter.com/yoheinakajima">@yoheinakajima</a> creator of babyAGI. Read their original thread here: <a href="https://twitter.com/yoheinakajima/status/1781470851621372238?s=48">https://twitter.com/yoheinakajima/status/1781470851621372238?s=48</a></figcaption></figure><p>Looking at this graph of Impact vs Cost, we discovered that businesses have been having significant problems with Zone 3 and Zone 2. These tasks often need to be completed on a daily or weekly basis and involve extracting large amounts of unstructured information from contractual documents, invoices, and emails and then entering that data into downstream systems such as CRMs and CMSs’.</p><p>Zone 3 represents relatively low-impact tasks that don’t directly affect one’s revenue, but have to be completed regularly with a high cost of doing so. Often this takes the form of bank reconciliations or invoicing. Depending on the size of the business this work sometimes needs a full team assigned to it to ensure that the sheer volume of data is processed in time. In smaller cases this can be a person’s full-time job, easily spending 20+ hours just on this task instead of other more productive activities related to the job description.</p><p>Moving to Zone 2, these are critical tasks that affect multiple areas of the business and require significant due diligence to avoid errors that could derail downstream tasks. This work might require multiple senior leaders’ full attention, and its cost can be measured in the time taken to complete it; either due to the volume of work or the input required from other sources of data.</p><p>Even more pressing, for both Zones 2 &amp; 3, is the opportunity cost relating to this work. Time spent completing these tasks is time taken away from other activities that could be growing the business and finding new opportunities for savings or revenue generation.</p><p>For completeness, Zone 1 represents tasks that are usually carried out once, or quite sporadically, usually of strategic importance. Think UX personas, competitor analysis, or roadmap planning.</p><h3>Positioning your LLM Agent</h3><p>There are numerous agentic startups (QoQo.ai, Taskade, Devin) that have positioned their product as part of a growing productivity toolset to be used in conjunction with business software such as Microsoft Word, Outlook, and even more specialized tools like Jira and Figma. While these tools go some way to alleviate the pain, often they solve problems in Zones 1 &amp; 3 and fall short of the type of solutions that could make a difference in Zone 2.</p><p>The aforementioned agent products focus on generating an end result, such as UX personas, support tickets, and technical specifications. While they have the potential to provide immense value, they have yet to address the problems described above. Instead, businesses need tools that can help them analyze their existing data, extract relevant details, and finally enter those into other systems.</p><blockquote>As a builder, your most valuable asset is the subject matter expertise you bring to the table for the particular niche your agent occupies.</blockquote><p>For instance, consider the following decisions to make when building a news summarising agent:</p><ol><li>Which news sources should you use?</li><li>Which elements of the news articles are particularly relevant?</li><li>How should that summary be used, or combined with other software systems?</li></ol><p>The correct answers to all these choices change depending on your customer, sometimes even based on the subcategory. Think real estate vs finance, private equity vs VCs, growth funds vs seed funds; the list goes on.</p><p>It’s on the next step where the builders that we’ve spoken to get stuck, however. While agents present a neat solution with the ability to shrink a 3-hour process down to 3 seconds, the following areas prove difficult:</p><ol><li>Obtaining customers who understand the value proposition of an agent for their organization</li><li>Identifying the right use cases across different industries where the agent excels at providing value</li><li>Building a common experience that can be adapted to work with each customer’s unique internal tooling environment</li></ol><p>Building an agent that satisfies a customer’s expectations for robustness, reliability, and consistency is already hard enough without the need to satisfy the above to build a sustainable business. Not to mention additional requirements around enterprise-style features that customers often request including data residency, multi-user access (RBAC), notifications, and alerting.</p><p>Recalling Zone 2 in the previous diagram, the most valuable problems within businesses involve sensitive data that are present in their internal tools such as Outlook for email, Salesforce for CRM, and SAP or Microsoft Dynamics for their ERP to name a few. This raises further questions:</p><ol><li>How should agents securely connect to these tools?</li><li>Which LLM API providers are compatible with the customer’s data locality requirements?</li><li>How and where should the customer’s data be stored, before and after it is processed by the agent?</li></ol><p>These are all non-trivial engineering problems that need to be addressed per locality to successfully serve a consistent experience to many different kinds of customers.</p><p>This brings us to the Solitude platform. We want builders to focus on what they do best. Builders should focus on developing a deep understanding of the customer problems within the verticals they want to address, and then selecting the right combination of technologies and algorithms to power the agentic solution.</p><p>There is an immense amount of value in solving these problems which are unique to each business. Solitude takes care of the common elements beneath them like infrastructure, data handling, and secure application access on behalf of your agent for tools like SAP, Outlook, Gmail, and Hubspot.</p><h3>Tradeoffs to make when building LLM agents</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dR5Jv82WV_MqpKdaKnkpQw.png" /><figcaption>A sneak preview of the Solitude Agent Builder</figcaption></figure><p>Similar to traditional software, each combination of technologies used to build an agent comes with its tradeoffs. When building an agent, consider the following constraints:</p><h4>LLM Agent Memory</h4><p>How should the agent’s memories be stored?</p><ol><li>GraphDB</li><li>RDMS</li><li>In-Memory Cache</li></ol><p>For how long?</p><ol><li>Time-based schemes</li><li>Priority-based schemes</li></ol><p>How do you tell the difference between useful memories and the rest?</p><ol><li>Relevance vs Importance vs Recency</li></ol><h4>LLM Agent Planning</h4><p>Which planning algorithm should you use?</p><ol><li>Reasoning as Planning</li><li>Reflection</li></ol><p>How complex does the plan have to be?</p><ol><li>Linear plans (ReAct)</li><li>Tree-based planning (Tree-of-thought)</li></ol><p>3. Graph-based planning (Graph-of-thought)</p><p>Will this impact the latency requirements for the customer?</p><ol><li>“Deep” planning requires several cycles, this usually incurs a ~30s response time (such as multi-level trees).</li><li>“Short” planning, which may be one-shot or two-shot plans, has a much faster ~1–3s response time (a simple ReAct execution pattern).</li></ol><h4>LLM Agent Tools</h4><ol><li>Which tools should the agent have access to?</li><li>What actions is the agent allowed to take on this tool? Can they read, write, or do both?</li><li>Will an RAG system be required to gather the correct inputs for that tool?</li></ol><p>While Solitude lets you configure all these options, it is up to you to decide what is best for the problems you’re trying to solve; which largely depends on the feedback you get from customers.</p><p>So you’ve built your agent, what next? In our next blog series, we’ll dive into how Solitude promotes your agent through our marketplace and other distribution channels to help buyers find your solution. We take care of details like payments and transaction so that you can get paid monthly for the work your agent carries out.</p><p>If you found this compelling, we encourage you to join our early access program where you can stay up to date with out progress and get an early invite to our platform: <a href="https://www.solitude.ai/">https://www.solitude.ai/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2ba054dfa17c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/how-to-build-a-llm-agent-for-enterprises-2ba054dfa17c">How to build a LLM agent for Enterprises</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Agent Marketplace: what is the end state of Generative AI?]]></title>
            <link>https://medium.com/solitude-agents/the-agent-marketplace-what-is-the-end-state-of-generative-ai-ce1dd9a57f31?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/ce1dd9a57f31</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <category><![CDATA[llm-agent]]></category>
            <category><![CDATA[enterprise-software]]></category>
            <category><![CDATA[generative-ai-market]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Mon, 15 Apr 2024 23:33:54 GMT</pubDate>
            <atom:updated>2024-04-16T23:46:54.794Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*B1-iumAEdSjNilNK" /><figcaption>A marketplace, one of the oldest structures in business. Source: <a href="https://unsplash.com/photos/aerial-photography-of-colorful-tent-M2524ncJQ40">https://unsplash.com/photos/aerial-photography-of-colorful-tent-M2524ncJQ40</a></figcaption></figure><p>If your business doesn’t find a cost effective way to implement AI, you could be losing millions of dollars that were spent on 100s of hours of manual data entry and missed opportunities as a consequence. For years, companies have been sold the idea that artificial intelligence (AI) will revolutionize their operations, enhance productivity, and drive innovation. Yet, the reality has been far less impressive. The adoption of AI has been slow, with many businesses struggling to integrate these technologies into their existing workflows and realize their full potential. The high costs, lack of expertise, and complexity of AI implementations have created significant barriers to entry, preventing many organizations from harnessing the power of AI. But what if there was a way to make AI more accessible, cost-effective, and easier to implement? Enter generative AI agents and the concept of an AI marketplace.</p><h4><strong>The Evolution of Generative AI in Business</strong></h4><p>The journey of generative AI in the enterprise has been one of cautious experimentation and gradual adoption. According to Census Bureau data, only about 5% of U.S. companies currently use any form of AI, with another 7% planning to adopt it in the next six months. This slow uptake can be attributed to several factors, including a lack of AI expertise, data silo issues, and concerns over IP leaks and public AI mistakes.</p><p>Despite these challenges, larger companies with dedicated AI teams and infrastructure investments are responsible for AI adoption. Industry giants like JPMorgan Chase and Bayer are experimenting with hundreds of use cases for generative AI agents, ranging from routine task automation to more advanced legal contract analysis and software development applications. However, much of this adoption is still surface-level, and the true transformative potential of generative AI agents still needs to be explored.</p><h4><strong>The Promise of an AI Marketplace for Generative AI Agents</strong></h4><p>To fully realize the potential of generative AI agents, businesses need a more accessible and cost-effective way to adopt and integrate these technologies into their operations. This is where the concept of an AI marketplace comes in. An AI marketplace would function as an “AWS for AI,” allowing companies to rent pre-trained, ready-to-deploy generative AI agents for various business functions on a pay-per-use basis.</p><p>Compared to other business models, such as building bespoke AI systems in-house or purchasing off-the-shelf AI solutions, an AI marketplace offers several advantages for deploying generative AI agents:</p><ol><li>Lower costs: An AI marketplace eliminates the need for significant upfront investments in AI infrastructure and expertise by providing generative AI agents as a service.</li><li>Faster time-to-value: With pre-trained generative AI agents available on-demand, businesses can quickly experiment and find the right AI fit for their needs without lengthy development cycles.</li><li>Scalability: An AI marketplace allows businesses to quickly increase or decrease their usage of generative AI agents based on their changing requirements.</li><li>Accessibility: By democratizing access to generative AI agents, an AI marketplace enables small and mid-sized firms to leverage powerful AI capabilities that were previously out of reach.</li></ol><h4><strong>The Road Ahead for Generative AI Agents and the AI Marketplace</strong></h4><p>The emergence of an accessible AI marketplace for generative AI agents is not an inevitability but a vision that requires aligning several economic and technical factors. On the financial front, there needs to be sufficient demand from businesses for generative AI agents as a service and a willingness to adopt new technologies and rethink existing workflows. This demand will be driven by the increasing pressure to stay competitive in an AI-driven world and the growing evidence of the transformative potential of generative AI agents across industries.</p><p>On the technical side, the infrastructure and ecosystem supporting an AI marketplace for generative AI agents must mature and evolve. This includes advancements in natural language processing, machine learning, AI architectures, and computing power and storage improvements. The development of standardized APIs and interfaces for seamlessly integrating generative AI agents into business workflows will also be crucial. As these technical foundations fall into place, the barriers to entry for adopting generative AI agents will continue to lower, paving the way for widespread usage.</p><h4><strong>Conclusion</strong></h4><p>The emergence of an accessible AI marketplace for generative AI agents can be a game-changer for businesses worldwide. It could unlock the transformative power of these technologies and usher in a new era of innovation and productivity. By providing a cost-effective, scalable, and accessible way to adopt generative AI agents, an AI marketplace can democratize access to these capabilities and enable businesses of all sizes to harness their potential.</p><p>As the economic and technical factors align to support the growth of an AI marketplace for generative AI agents, businesses that embrace this new paradigm will be well-positioned to thrive in the AI-driven future. The road ahead may be challenging, but the destination — a world where the power of generative AI agents is accessible to all — is well worth the journey.</p><p>If you want to be ahead of the curve and want to get involved in the first Agent Marketplace, you can find out more here: <a href="https://www.solitude.ai/">https://solitude.ai</a></p><h4>Sources</h4><ol><li><a href="https://www.economist.com/the-world-ahead/2023/11/13/generative-ai-will-go-mainstream-in-2024">https://www.economist.com/the-world-ahead/2023/11/13/generative-ai-will-go-mainstream-in-2024</a></li><li><a href="https://www.economist.com/business/2024/02/29/how-businesses-are-actually-using-generative-ai">https://www.economist.com/business/2024/02/29/how-businesses-are-actually-using-generative-ai</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ce1dd9a57f31" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/the-agent-marketplace-what-is-the-end-state-of-generative-ai-ce1dd9a57f31">The Agent Marketplace: what is the end state of Generative AI?</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The rise of the AI employee]]></title>
            <link>https://medium.com/solitude-agents/the-rise-of-the-ai-employee-c5712221f8e2?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/c5712221f8e2</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <category><![CDATA[enterprise-automation]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Mon, 08 Apr 2024 23:48:24 GMT</pubDate>
            <atom:updated>2024-04-09T13:20:58.423Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nlKPXhcpPBQqey4U.jpg" /><figcaption>The BD-1 droid from the Book of Boba Fett. Source: <a href="https://www.slashfilm.com/img/gallery/the-book-of-boba-fett-brings-back-a-familiar-adorable-droid-model-from-star-wars-jedi-fallen-order/l-intro-1643237682.jpg">https://www.slashfilm.com/img/gallery/the-book-of-boba-fett-brings-back-a-familiar-adorable-droid-model-from-star-wars-jedi-fallen-order/l-intro-1643237682.jpg</a></figcaption></figure><p>How would your business change if it was run by 1000 AI employees for the cost of a single entry-level position?</p><p>Sequoia Capital recently wrapped up its AI Ascent summit, where they brought in AI leaders from the most successful early-stage startups and public companies around the world to discuss the future of the industry. Even esteemed AI leaders aren’t too sure how the current wave of technology will play out. Although they all agree on one thing; generative AI agents represent the next huge sea change in technology on the scale of mobile phones and the internet. In this article, we’ll zoom in on a few talks that our team found particularly illuminating which struck a chord with our customers (read more <a href="https://medium.com/solitude-agents/what-is-an-llm-agent-for-non-techical-people-2baec0326583">here</a> to learn about agents in more detail).</p><p>Sequoia’s opening keynote set the stage for how they perceive the next 10 years will play out for businesses that make AI a cornerstone of their operation. They predict that traditionally capital-intensive incumbents in industries like finance and healthcare will leverage agents to massively reduce marginal costs. This would produce a downward pressure on prices to bring them in line with products in computing and software which were one of the only industries that has gotten cheaper for the consumer over time. New entrants to the market that are building AI-based businesses in these areas will have a headstart without the need to migrate existing systems to the “agentic” way of doing things. They express that the true potential is only unlocked when these agents are directly linked to business KPIs that allow them to drive the status quo of a business by incorporating them into their feedback loop for completely autonomous, iterative improvement (read more <a href="https://medium.com/solitude-agents/what-is-an-llm-agent-for-non-techical-people-2baec0326583">here</a> if you want to learn what an agent is).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OS_KyFw7vlJ7039q5GxAag.png" /><figcaption>Screenshots from Sequoia’s opening keynote</figcaption></figure><p>The overall theme is clear, a confluence of events is forcing businesses to adapt and change to this new technological climate, the first wave was LLMs and the next will be agents. A business organization that’s operated by 1000 agents never has to worry about the increasing complexity of communication and alignment that a traditional business has to deal with as it scales operations. More often than not, the greatest obstacle to innovation in established businesses is <em>figuring out where your data is and then finding opportunities for innovation.</em> With agents, this would never be a problem as they always have all the data they need; they can instantly offer insights about areas of the business that could be improved or optimized.</p><h4>Can AI agent technology make the jump to enterprise?</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*E1sUrGBDeftsFSQw.jpg" /><figcaption>The Apple Lisa and its accompanying GUI. Source: <a href="https://macgui.com/upload/gallery/f_0/user_2/regular/upload_5179.jpg">https://macgui.com/upload/gallery/f_0/user_2/regular/upload_5179.jpg</a></figcaption></figure><p>This poses an immediate question: Is this even possible? CJ Desai, COO of the now famous enterprise automation platform ServiceNow made it clear that AI has already been making an impact for his customers. As a big supporter of open-source LLMs, they’ve put them to work by massively reducing the time it takes to complete tasks like inventory management or invoice reconciliation. For large enterprises, open source models are much more capital efficient which can make a big difference if a 1% change in margin can make or break your quarter. He sings to the same tune as Andrej Kaparthy (Director of AI at Tesla), who agreed that the open source ecosystem has vast amounts of untapped potential, but he still thinks there’s significant room for improvement: “You build a product, then a ramp. But [people] are just building products with no ramp!” he exclaimed. A product with no ramp draws on images of computers before the Apple Lisa. Without a simple point-and-click graphical user interface, computers at the time were relegated to purely technical audiences, despite the huge uplift in productivity they came to embody in the years to come. The added layer of accessibility that Apple introduced suddenly turned the computer into a savvy tool anyone could start with and ramp up to more advanced use cases.</p><p>Andrej goes on to point out a big gap between the capabilities AI agents have showcased in tech demos and agents that can produce a real-world impact on your business. For example: creating content that’s consistent with your brand voice to drive engagement across your go-to-market channels. Consistency is critical when it comes to operating agents for these use cases at scale. Fortunately, this technology has been evolving at an extremely rapid pace, and guardrails for this consistency have already been engineered and put to work. One of the greatest advancements here is agent planning, which effectively allows agents to document the actions they intend to take so that they can be corrected and or approved.</p><h4>From LLMs to AI agents</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AdSxFE8QxQkJhXHNGK9LYw.png" /><figcaption>Andrew Ng presenting planning algorithms</figcaption></figure><p>“For those who haven’t experienced agents that plan [ahead], it’s a real <em>aha</em> moment on the level of that first experience with ChatGPT,” remarks Andrew Eg, co-founder of Coursera and Professor of Machine Learning at Stanford. He echoed the message in the opening keynote: As AI agents can automate more and more of the workplace, we’ll begin to see entire organizations that just operate as a unified team of agents, constantly working to optimize and improve the KPIs that drive your business. It remains to be seen if the technology we use today will become established as the go-to method or if we’ll invent new ones along the way. What’s for certain is that businesses need to start taking a serious look at how AI will impact their strategy in the next 12 months and start finding the easiest ways to test the waters; that’s where Solitude has you covered.</p><p>We want to make running 1000s of agents for your business simpler than your morning email routine. As new AI use cases emerge and agents become commonplace, managing them and orchestrating their work will become astronomically difficult, especially if their services and billing are spread out across 10–20 different vendors. If Sequoia’s vision plays out, this will be the least of the problems when dealing with agents sprawled across so many different platforms. What about governance and security? You shouldn’t have to worry about all this as running your business is already hard enough. We believe that managing and orchestrating fleets of agents should be as easy and intuitive as any other software you use today; point, click, complete. Other leaders in the space seem to agree with us too, all of whom reference the promising results that enterprises have been achieving today while acknowledging the shortfall in the UX of existing products.</p><p>“UX for agents is underbaked” Harrison Chase sighs. As one of the creators of Langchain, a technology that underpins agents, he agrees that AI agents are here to stay and will change the way that we work even though they present a serious challenge in terms of usability. “Chat boxes are insufficient” and if you’re an enterprise trying to solve real-world problems then the truth of this statement is immediately apparent. We too discovered that chat boxes are not suitable for scaling enterprise workflows because human users need to have insight and control of the actions the agents are taking. As a decision maker, understanding the thought process and reasoning that went into any content or output the agent generates before it is actioned is critical. Agent notifications and human approval workflows keep you constantly informed of key events and outcomes that the agent was responsible for while ensuring you have the final say on any action the agent pursues. For our customers, access to a central dashboard where they can freely view and approve any requested actions by agents is critical. This helped them avoid the flood of emails and notifications from other systems that can quickly become white noise.</p><p>The future of agents for the enterprise is extremely promising when you look at the vast range of use cases they can deliver across marketing, sales, customer service, and even software development and finance. With 1000s of agents at your fingertips, 1-person businesses that breach $100m+ ARR are not too far off. The technology is here today, but key issues around management and governance must be addressed first before we can confidently put them into production, which is why we built Solitude with those core principles in mind.</p><p>If you’re curious for a sneak peek at some of the platform features we mentioned today, our website is a great place to start: <a href="https://www.solitude.ai/">https://www.solitude.ai/</a></p><h4>Sources</h4><ul><li>Opening Keynote</li><li><a href="https://www.youtube.com/@sequoiacapital">https://www.youtube.com/@sequoiacapital</a></li><li>Harrison Chase</li><li><a href="https://www.youtube.com/watch?v=pBBe1pk8hf4&amp;ab_channel=SequoiaCapital">https://www.youtube.com/watch?v=pBBe1pk8hf4&amp;ab_channel=SequoiaCapital</a></li><li>Andrew Ng</li><li><a href="https://www.youtube.com/watch?v=sal78ACtGTc&amp;ab_channel=SequoiaCapital">https://www.youtube.com/watch?v=sal78ACtGTc&amp;ab_channel=SequoiaCapital</a></li><li>Andrej Karpathy</li><li><a href="https://www.youtube.com/watch?v=c3b-JASoPi0&amp;ab_channel=SequoiaCapital">https://www.youtube.com/watch?v=c3b-JASoPi0&amp;ab_channel=SequoiaCapital</a></li><li>CJ Desai</li><li><a href="https://www.youtube.com/watch?v=-hUJTEDOQDs&amp;ab_channel=SequoiaCapital">https://www.youtube.com/watch?v=-hUJTEDOQDs&amp;ab_channel=SequoiaCapital</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c5712221f8e2" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/the-rise-of-the-ai-employee-c5712221f8e2">The rise of the AI employee</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From chatbots to AI Agents]]></title>
            <link>https://medium.com/solitude-agents/from-chatbots-to-ai-agents-ddfcabf6e8bc?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/ddfcabf6e8bc</guid>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <category><![CDATA[automation]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Tue, 02 Apr 2024 23:22:38 GMT</pubDate>
            <atom:updated>2024-04-06T13:30:49.411Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rwM_E_3spA2zGzs9rI3Jgg.jpeg" /><figcaption>From the creative minds here: <a href="https://deerdesigner.com/blog/the-pros-and-cons-of-letting-ai-take-the-design-reins/">https://deerdesigner.com/blog/the-pros-and-cons-of-letting-ai-take-the-design-reins/</a></figcaption></figure><p>The adoption of AI agents has faced significant challenges in the enterprise. Businesses face hurdles such as the need for internal expertise, data complexity, and potential risks like intellectual property leakage. Skepticism about AI’s immediate productivity benefits and implications for job quality and output can also hinder further adoption.</p><p>Generative AI has emerged as a transformative technology that reshapes how businesses operate and interact with customers. To harness this technology effectively, it’s essential to understand current applications, limitations, and future potential to grasp how it will impact the business landscape in the coming years.</p><h4><strong>How did business get started with AI?</strong></h4><p>The journey of generative AI in business began with early adopters of Natural Language Processing (NLP). This was one of the first areas where generative AI made an impact, with businesses using it for content creation, translation, and sentiment analysis. Companies also started exploring the potential of generative AI in image and video creation, using it for marketing, design, and entertainment purposes. Generative AI also found its way into data augmentation and synthetic data generation, helping businesses train machine learning models more effectively.</p><p>As generative AI technologies matured, businesses began incorporating them into customer-facing applications like chatbots. These AI-powered conversational interfaces have become increasingly common, allowing companies to provide round-the-clock customer support and streamline routine interactions. Chatbots have delivered mixed results, with the limited ability to handle basic interactions but still relying on a human-in-the-loop to fulfil more complex requests. However, the recent advancements around large language models (LLMs) challenge what we thought the limits of automation were.</p><h4><strong>What new generative AI technology is on the horizon for businesses?</strong></h4><p>What if you could hire an employee who never slept, worked 24/7, and never complained or filed sick? Short of a few human rights violations, the next best thing would be this quickly evolving technology known as AI agents. AI agents are the successor to the LLM chatbots we’ve seen today. They can automate complex tasks that require reasoning, enhance decision-making, and enable new forms of value creation.</p><p>AI agents are more versatile and adaptable than chatbots which were designed for specific, narrow tasks. Agents can be trained to perform various functions across various industries, from content creation and data analysis to process automation and decision support. As businesses begin to recognize the potential of AI agents, we are seeing a surge in investment and experimentation.</p><p>To describe the overall effect this upcoming wave of AI agents could have on businesses, we refer to “Total factor productivity” (TFP). TFP is a recent term coined by Tinbergen (1942) that describes increases in economic output that cannot be attributed to measurable inputs, such as capital and human labor. Instead, TFP is driven by technological advancements, process improvements, and other intangible factors that enable businesses to produce more with the same amount of inputs. AI agents represent a unique technology with the potential to unlock unprecedented levels of innovation across industries and significantly contribute to TFP growth.</p><h4><strong>Is now the right time for enterprises to jump into Generative AI?</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/678/1*0ogheU3qHCCco-lqI85wsw.jpeg" /></figure><p>The market impact of generative AI has been substantial. Since the release of GPT-4, the U.S. technology industry’s market capitalization has risen by 50%, adding $6 trillion in shareholder value. Companies like Nvidia, Alphabet, Amazon, and Microsoft have seen their share prices soar due to their investments in AI. However, despite the hype, AI software sales are still relatively modest, accounting for only a tiny portion of the revenue growth at major tech companies. This suggests that the sustainability of the AI stock market boom will depend on the broad commercial application of AI technologies.</p><p>Adoption of generative AI is still in its early stages, with only about 5% of American firms currently using some form of AI and an additional 7% planning to adopt it within six months. However, large corporations are already exploring the potential of AI agents, with companies like JPMorgan Chase, Capgemini, and Bayer boasting hundreds of AI use cases across various functions. These applications range from simple tasks for low-skilled workers to complex tools for highly skilled employees, showcasing the versatility of AI agents.</p><p>Despite the current challenges and cautious adoption, early evidence suggests that AI might be more about job creation than elimination. Only 12% of corporations view AI as a replacement for human labour, and new job categories like “prompt engineers” and “explainability engineers” are emerging. As businesses continue experimenting with generative AI, we see practical applications across industries, from customer service improvements and administrative efficiency to enhanced research and data processing.</p><p>If your business has hit a roadblock in productivity or automation, then it might be time to start investigating what generative AI solutions can do for you before your competitors get there first.</p><h4><strong>Can businesses start using AI agents today?</strong></h4><p>Businesses that embrace AI agents and adapt to this new paradigm of productivity will be best positioned to thrive in the era of generative AI and become a TFP growth leader in their industry. The only question is, where can businesses go to start trying these technologies without a huge upfront investment?</p><p>As the market expands and vendors offer increasingly niche products, it becomes harder and harder for non-technical decision makers to search and find AI products that can truly shift the dial for their business. As of today, businesses have no way of easily onboarding this advanced technology without a significant amount of in-house investment and expertise. Even selecting a vendor presents challenges; the understanding required to compare and contrast different vendors on their merits is largely technical.</p><p>Solitude aims to provide a centralized hub where businesses can discover, test, and implement AI agents specific to their needs. By offering a curated selection of high-quality AI products and a user-friendly interface, Solitude helps businesses find AI Agents for their use cases built by expert sellers all in one place; enabling them to compare the quality of different agents without a PhD!</p><p>As always, if you’ve enjoyed reading this and want to learning more about how Solitude can supercharge your business; we welcome you to join our waitlist for an invite to our early adopter’s program: <a href="https://solitude.ai/">https://solitude.ai</a></p><h3>Sources</h3><ol><li><a href="https://www.economist.com/business/2024/02/29/how-businesses-are-actually-using-generative-ai">https://www.economist.com/business/2024/02/29/how-businesses-are-actually-using-generative-ai</a></li><li>“AI Adoption in the Enterprise 2023” by O’Reilly (<a href="https://www.oreilly.com/radar/ai-adoption-in-the-enterprise-2023/">https://www.oreilly.com/radar/ai-adoption-in-the-enterprise-2023/</a>)</li><li>“Generative AI Set To Revolutionize Industries And Propel The U.S. Economy” by Forbes (<a href="https://www.forbes.com/sites/nishatalagala/2023/03/31/generative-ai-set-to-revolutionize-industries-and-propel-the-us-economy/">https://www.forbes.com/sites/nishatalagala/2023/03/31/generative-ai-set-to-revolutionize-industries-and-propel-the-us-economy/</a>)</li><li>“The Rise of Generative AI” by Harvard Business Review (<a href="https://hbr.org/2021/07/the-rise-of-generative-ai">https://hbr.org/2021/07/the-rise-of-generative-ai</a>)</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ddfcabf6e8bc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/from-chatbots-to-ai-agents-ddfcabf6e8bc">From chatbots to AI Agents</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is an LLM Agent for non-techical people?]]></title>
            <link>https://medium.com/solitude-agents/what-is-an-llm-agent-for-non-techical-people-2baec0326583?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/2baec0326583</guid>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[llm-agent]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <category><![CDATA[enterprise]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Tue, 26 Mar 2024 03:44:15 GMT</pubDate>
            <atom:updated>2024-03-27T02:03:58.556Z</atom:updated>
            <content:encoded><![CDATA[<h3>What is an LLM Agent for non-technical people?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gAgBJi3tRwf0AquU6yj6Dg.jpeg" /><figcaption>The Baymax “Agent” from Big hero 6</figcaption></figure><p>In our ongoing to series to discover what the future of automation might look like, today we’re taking a deep dive on Agents for all you non-techies out there. Even though this subject area can get technical very fast, we wanted to present a simple, easy to understand introduction to LLM agents.</p><p>First, some basics. An LLM (Large Language Model) is a form of AI that uses natural language processing techniques to generate meaningful text that often comes in the form of dialog. This is the engine that powers ChatGPT.</p><p>Agents are an extremely new development at the forefront of this technology which represents the next stage of the LLM capabilities. As users, we’ve become accustomed to prompting the LLM with some text to generate an appropriate response. Agents take things a step further and allow the LLM to take actions on behalf of the user (e.g. schedule a meeting for you, like Siri or Alexa).</p><p>It’s easier to start explaining what an Agent is by first listing a few of the things it is not.</p><p>An Agent is not:</p><ul><li>A chatbot</li><li>A simple LLM (by itself)</li><li>A person (yes really)</li></ul><p>Agents are a combination of advanced programming and prompting techniques which allow an LLM to reason about the world around it including any context you give it, and take actions to accomplish goals you set for it. It’s like an advanced version of Siri or Alexa that can not only set meetings when you tell it to, but also engage in fluid conversation and proactively manage your calendar, alter spreadsheets, interact with software, and generate content in response to the conversation you’re having with it.</p><p>It’s best to understand an Agent’s capabilities with a few examples.</p><p>Take the following instruction:</p><blockquote>“Book a meeting tomorrow at 09:00 if the client is available, if she cancels send her a follow-up email and update the records in our CRM.”</blockquote><p>The following will occur:</p><ul><li>The agent will understand from your conversation history that the “client” you are referring to is the (fictional) email `<a href="/service/mailto:jane@my-client.net">jane@my-client.net</a>`</li><li>The agent will then check Jane’s calendar and see if there is a slot open for 09:00 and try to book a meeting.</li><li>If Jane cancels, the agent will draft a full follow-up email trying to book another date</li><li>Finally, the agent will update the details of the expected meeting in your CRM.</li></ul><p>The tasks you give the agent can also be even more abstract:</p><blockquote>“Develop a content strategy based on the best-performing posts by engagement across our LinkedIn, X, and Instagram channels for the next 2 months, it must include influencer marketing”</blockquote><p>In this case, the agent would go ahead and read all the metrics across your media channels and identify those with the highest engagement. It would then use those to form a base from which it builds a content strategy that it lays out in Excel (or Google Sheets, we don’t judge). Finally, based on the posts and your brand voice, it would search the web for the best influencers to partner with as part of the strategy.</p><p>You didn’t tell it to search the web, it reasoned that it had to do so (in the absence of specific instructions) in order to identify the influencers that would be right for you. You get the picture.</p><p>With the examples out of the way, we can define all the ingredients for our lovely agent soup. An Agent is an LLM augmented with the following capabilities:</p><ol><li>Long term memory</li><li>Tools to use (e.g a calendar or a web browser)</li><li>Logical reasoning</li><li>Planning</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MObNwFdbMOcBZbV1Kvwl8g.png" /><figcaption>A high level diagram of Agents interacting with Humans</figcaption></figure><h4>Why is an Agent different from a regular LLM?</h4><p>A regular LLM doesn’t have the capabilities listed above, namely long-term memory, planning, tools, and reasoning. An LLM has a context window (a maximum word count for its inputs and outputs) and that’s the extent of its memory! Agents on the other hand significantly extend what’s possible by augmenting the LLM with supporting technology.</p><p>LLMs also cannot use tools like regular software or web browsers, because it has no way to interact with the environment or produce an output other than text. Agents can engage in conversation, extract what’s relevant, and use that as input into other software.</p><p>Moreover, an Agent can be “taught” to use any tool! Whether it’s a web browser, Outlook, Google Calendar, or your favourite messaging application like Slack, Teams, Whatsapp, or Telegram.</p><p>Reasoning is what ties this all together. The agent will attempt to reason about the task being asked of it by chaining together “thoughts”. Without getting technical, “thoughts” enable the LLM to inspect its own output and decide if it is relevant or useful to the task at hand.</p><p>Planning is a fancy word for letting the LLM “think” ahead and decide on a series of steps it will take to accomplish the goal, instead of focusing only on the immediate next step.</p><h4>Agents are the next enterprise platform</h4><p>Given the amazing capabilities Agents have demonstrated we believe they will fundamentally change the way we work in the future. As shown in the content strategy example above, Agents are a first-of-a-kind technology with the ability to autonomously complete complex tasks from end to end. Although the technology is still nascent, it has the potential to change the way we think about work and coordinate large-scale knowledge-based work in the future.</p><p>That said, there are still some limitations we came across when building these in-house. We found that for the best performance, an Agent should specialise in a given domain (e.g. Marketing, Sales); even better if it can niche down on a vertical like Finance Marketing or Tech Sales.</p><p>Just like people, specialisation can help ensure that the Agent produces results with a consistent quality that’s perfect for your business. More generalised Agents can get the job done but can be unspecific in their approach, this can affect the quality of any generated content. Just like ChatGPT, non-specialised Agents can get by in terms of general knowledge, but their performance drops off when you start to ask it for specific details which might be niche to your business.</p><h4>How can I use LLM Agents in my business?</h4><p>If you’re asking yourself this, you’ve come to right place!</p><p>At Solitude, we’re building a platform to connect Agent builders with innovate enterprises who are looking to automate more of their business than ever before. For enterprises, you can think of the Solitude as Upwork or Fiverr for AI employees that can carry out any task, which you’re free to hire on a monthly basis.</p><p>Builders, can use our platform tools to develop specialised agents for any job role including Marketing, Sales, Customer Service and Finance to name a few. Agents can build for any functional area, including ones we have listed here like Design!</p><p>If you found this an interesting read and want to learn more, we’d love to extend an invite to you via our early adopter’s program which you can sign up for at our website: <a href="https://solitude.com/">https://solitude.com</a></p><h4>Sources</h4><ol><li><a href="https://lilianweng.github.io/posts/2023-06-23-agent/">https://lilianweng.github.io/posts/2023-06-23-agent/</a> — for more technical audiences</li><li><a href="https://www.forbes.com/sites/alexanderpuutio/2024/03/22/what-ceos-need-to-know-about-the-next-frontier-of-ai-ai-agents/">https://www.forbes.com/sites/alexanderpuutio/2024/03/22/what-ceos-need-to-know-about-the-next-frontier-of-ai-ai-agents/</a> — a Forbes article on agents</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2baec0326583" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/what-is-an-llm-agent-for-non-techical-people-2baec0326583">What is an LLM Agent for non-techical people?</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[UX for AI is broken, and we need to re-think it from scratch]]></title>
            <link>https://medium.com/solitude-agents/ux-for-ai-is-broken-and-we-need-to-re-think-it-from-scratch-900b78da28e5?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/900b78da28e5</guid>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[generative-ai-consulting]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[history-of-technology]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Tue, 19 Mar 2024 00:40:56 GMT</pubDate>
            <atom:updated>2024-03-19T00:40:36.046Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vzIIqKUT1QeuiEuQp2iI6w.png" /><figcaption>From an awesome source you should checkout: <a href="https://www.futurity.org/robots-mistakes-trust-2868702/">https://www.futurity.org/robots-mistakes-trust-2868702/</a></figcaption></figure><p>Beyond the simple chat interfaces that have become the de-facto standard for today’s AI products, business users are struggling to get these highly advanced models to do what they want, but it seems so simple! The impact of UX is often subtle and not immediately obvious, yet collectively, it significantly enhances the products utility. The adoption and popularity of a technology is intricately linked to how its users interact with the product, and their experiences can make or break a product’s success.</p><p>In this article, we’ll dive into the history of the tumultuous relationship between the main characters in this story: AI and UX.</p><h3>The Georgetown-IBM Experiment and the Onset of AI Winters</h3><p>The crux of AI’s developmental narrative has been profoundly influenced by cycles of inflated expectations and subsequent reality checks. While not exclusive to AI, this pattern is particularly striking within this field due to the ambitious promises associated with its technologies. A notable example that encapsulates this trend is the Georgetown-IBM experiment, which led to the creation of the ALPAC report.</p><p>The ALPAC report stands as a significant turning point in the history of AI, casting a long shadow over the potential of machine translation — an key early manifestation of an AI enabled product being successfully introduced to the market. The report’s critical view of the achievements in machine translation at the time was stark, pointing out a glaring mismatch between the hyped expectations and the actual capabilities of AI technologies. This discrepancy did not just draw attention to the limitations of AI at the time; it inadvertently heralded the onset of what is known as the first AI winter. This period was marked by a significant downturn in interest and funding for AI research, driven by the disillusionment with the field’s progress compared to the high expectations set.</p><p>This historical context points to a crucial problem: recurrent hype cycles in AI development lead to exaggerated capabilities, culminating in significant setbacks for the field. These cycles are not merely patterns of financial investment and academic interest but reflect more profound issues in how AI technologies are presented and perceived by the public, policymakers, and the research community. The mismatch between what AI promises and what it can deliver at various points in its evolution exacerbates the challenges of developing and integrating AI technologies into practical applications.</p><p>AI winters and subsequent hypercycles are intrinsically linked. Typically, a breakthrough generates high hopes, as in the case of machine translation in the 1990s. These initial successes fuel a surge in interest and investment, creating a hype cycle that elevates expectations beyond the current technological capabilities. When AI’s limitations become apparent, disillusionment, marked by reduced funding and interest, leads to an AI winter.</p><p>The problem of mismatched expectations versus capabilities is historical and remains relevant today as AI technologies evolve. Understanding this pattern is crucial to addressing the challenges of developing AI technologies that meet society’s practical needs while avoiding the pitfalls of past hype cycles.</p><h3>Navigating Through the Hype: The Role of UX in AI’s Evolution</h3><p>As AI navigated through periods of hype, the renaming of technologies emerged to circumvent the negative connotations associated with AI failures. Terms like “expert systems” and “neural networks” reflected a shift towards more specialized language, aiming to secure funding and support without invoking the baggage of past disappointments. This renaming underscored a broader issue: the misalignment between AI’s perceived and actual capabilities, a gap that UX design uniquely bridges.</p><p>The integration of AI into consumer applications in recent decades underpins the escalating expectations users have for technological interactions. The transition from a world with limited household technology to one that is inundated with digital devices and services has amplified the demand for intuitive, user-friendly experiences. This marks a turning point for AI’s place in the broader ecosystem — no longer an invisible technical feature but a source that drives meaningful human interactions. As AI becomes more embedded in everyday applications it is becoming increasingly clear that the UX layer that sits atop these technologies has become the point of differentiation, not the capabilities of the models themselves.</p><h3>The Renaissance of AI: Learning from the Past</h3><p>The current resurgence of AI, fuelled by groundbreaking advancements in machine learning and deep learning, marks a significant departure from its historical cycles of exaggerated expectations and subsequent disillusionments. This renaissance, characterized by a more refined understanding of AI’s capabilities and limitations, fosters an environment where AI can thrive in well-defined domains. Yet, the shadow of past AI winters looms, a stark reminder that neglecting user experience (UX) could propel us back into a cycle of stagnation.</p><p><strong>The Evolution of UX in AI</strong></p><p>In the past, AI’s journey through hype cycles often lacked a crucial component: a deep focus on user experience. This oversight contributed to the mismatch between AI’s promised capabilities and its actual utility, leading to periods of disillusionment. However, today’s AI renaissance is built on user-centric design principles that aim to rectify past mistakes.</p><ol><li><strong>Understanding User Needs: </strong>Early AI developments often prioritized technological innovation without thoroughly understanding end-user needs or how users interact with the system. This approach led to solutions that, while technologically advanced, needed to be aligned with real-world applications or user expectations. In contrast, modern AI developments leverage user research and feedback loops to deeply understand and anticipate user needs, ensuring that AI solutions are innovative but also practical and user-friendly.</li><li><strong>Iterative Design and Testing: </strong>Previously, an engineering team would develop the application and put it into the wild based on their own assumptions about the user, without any customer discovery or user testing. Today, iterative design processes are key to successful AI development. These processes involve continuous testing, feedback, and refinement, incorporating user input at every stage to ensure the final product is intuitive and effective. This iterative approach helps prevent the disconnect between AI’s capabilities and user expectations.</li><li><strong>Accessibility and Inclusivity</strong>: Modern UX practices in AI emphasize making technology accessible and inclusive, recognizing the diverse needs of all potential users. This shift towards designing AI applications usable by people with a wide range of abilities and in various contexts is a significant departure from past practices.Focusing on these elements,, AI developers can ensure their technologies provide value to a broader audience, further enhancing adoption and satisfaction.</li><li><strong>Transparency and Trust: </strong>One of the lessons learned from past AI winters is the importance of building trust with users. Modern UX strategies emphasize the importance of transparency in AI systems, making it easier for users to understand how AI makes decisions. This clarity helps build trust and demystifies AI technology, making users more comfortable with and receptive to AI solutions.</li></ol><h3>The way we build AI solutions has totally changed.</h3><p>Recent trends have significantly accelerated the pace at which firms can meet consumer demands. This acceleration is partly due to a shift in how AI solutions are developed, moving away from the traditional, opaque, and costly process that requires firms to hire expensive specialists or rely on AI consultancies. These conventional methods often suffered from misaligned incentives, as consultancies benefited from prolonged project timelines, and certainly not the efficient delivery of solutions that match the client’s needs.</p><p>In contrast, the current landscape is shaped by the rapid adoption of AI technologies. This evolution drives the market toward a solution that minimizes friction in building and utilizing AI tools. The market’s readiness for such solutions highlights the necessity for platforms that streamline the creation and deployment of generative AI solutions, making it easier and more cost-effective for businesses to access the latest innovations.</p><p>In our next blog, we will explore in detail how businesses can implement these user-centric practices. Stay tuned as we unravel the evolution of UX for AI products beyond the simple chat interfaces we see today.</p><p>As always, if you found this article insightful we would love to see you in our early adopter’s program where you can test a few of these things out for yourself! <a href="https://www.solitude.ai/">https://www.solitude.ai/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=900b78da28e5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/ux-for-ai-is-broken-and-we-need-to-re-think-it-from-scratch-900b78da28e5">UX for AI is broken, and we need to re-think it from scratch</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Is the automation industry stagnating?]]></title>
            <link>https://medium.com/solitude-agents/is-the-automation-industry-stagnating-bcf9d79dca18?source=rss----a1447ff6a896---4</link>
            <guid isPermaLink="false">https://medium.com/p/bcf9d79dca18</guid>
            <category><![CDATA[no-code]]></category>
            <category><![CDATA[rpa-tools]]></category>
            <category><![CDATA[automation]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[ai-agent]]></category>
            <dc:creator><![CDATA[Solitude]]></dc:creator>
            <pubDate>Tue, 12 Mar 2024 10:11:30 GMT</pubDate>
            <atom:updated>2024-03-20T14:00:10.531Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="A snapshot of the the Solitude drag and drop ai agent builder" src="https://cdn-images-1.medium.com/max/1024/1*i7mCsDm6aJqdbnVB1z2wzg.png" /><figcaption>A preview of the Solitude Agent Builder</figcaption></figure><p>Companies and enterprises alike have been spending loads on automation over the past decade, and that spending has taken many forms:</p><ul><li>RPA (Robotic Process Automation)</li><li>No Code / Low Code Software</li><li>Hyperautomation (RPA + AI)</li></ul><p>While this is great, where are all the gains we supposedly should have attained by way of this automation? We can point to any number of organisations that have published the savings they have made in time and possibly revenue by implementing these tools, but as an industry, we keep missing the mark when it comes to delivering on the promise splattered across every nascent automation company’s hero section: “less time doing boring stuff, more time doing work that matters”.</p><p>Despite the success on the surface, we took a deeper look on the forums where the people who actually have to build and implement the solutions for their company congregate: Reddit &amp; Twitter. As it turns out, the developers charged with building and maintaining these solutions all harbour common complaints:</p><p>1. So called “Citizen Developers” rarely create robust and useful automations<br>2. Most automations created are quite fragile and deeply coupled to the underlying process that they’re built around<br>3. The UX for developing these automations can be quite poor, especially in the case of market leading RPA solutions</p><p>So what gives? To understand this better, lets segment these tools by the market they address.</p><h3>Large enterprises often use RPAs for automation</h3><p>Large enterprises have software and processes that are very difficult to automate (e.g they have no API surface), or, the existing jungle of red tape, ERP systems and BAU management functions make finding what needs to be automated <em>extremely difficult</em>. In to the rescue comes RPA.</p><p>RPA solutions are unique because they are built specifically for these use cases; this software segment focuses on automating mouse movements and keyboard entries. Using an RPA tool goes something like this:</p><p>First, record yourself completing a given task, warts and all:</p><ol><li>Open your ERP,</li><li>Click around,</li><li>Enter some customer details,</li><li>Close it</li><li>Open Excel</li><li>Add some details</li><li>Finish.</li></ol><p>Then, play back this recorded routine on a set schedule.</p><p>I’ve glossed over the fine details here but it slowly becomes clear why this process is so fragile.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/590/1*i530Es13ASGA2wzsIEMuGg.jpeg" /><figcaption>Source: <a href="https://www.reddit.com/r/automationmaniac/comments/ymn5jr/one_man_army/">https://www.reddit.com/r/automationmaniac/comments/ymn5jr/one_man_army/</a></figcaption></figure><p>RPA works best when a process is fairly unchanging and extremely repetitive, any change to the sequence defined above would mean that you have to go in and <em>record the entire process all over again</em>. Since theres no API surface (with certain notable exceptions for legacy software, Oracle Developer Server I’m looking at you), this is probably the best solution for now.</p><p>That said, much of the frustration arises when the process inevitably changes. The existing work has to be reprogrammed, which is especially painful if it was originally put in place by “Citizen Developers” as they are usually not provided adequate training for building scalable automations that are robust to edge cases.</p><p>The gain in productivity is achieved in exchange for the friction of updating the automation over time. Not to mention, the cost of skilled workers or consultants assigned to the maintenance work. Should this skill and knowledge leave the organisation without capturing the information, the value of the automation is significantly diminished.</p><p>If an organisation evolves and needs to update their procedures, they will either have to replace the entire automation, or retrain new employees on the automation tooling to accommodate the required changes.</p><h3>Startups benefit from low/code no code tools to move faster</h3><p>In a completely different market segment, low code / no code shines where there are APIs openly available that can be used to connect various solutions. Ergo the myriad of integrations the market incumbents like Zapier and others offer.</p><p>These tools come in many different flavours, website builders, workflow automation and design-to-code tools to name a few; we’ll focus on workflow automation tools today.</p><p>At small-medium sized companies and especially startups that might not be around in the 24 months, there is a heightened urgency to remove any busy-work so that the team can focus on iterating the core product / service and achieving the holy-grail of product-market-fit. In this respect, using low code / no code tools to quickly automate business processes outside of the core business domain such as accounting, finance and and CRM flows, quickly demonstrates an ROI during the early lifecycle of the company. Over time however, these companies begin run into the same points of friction we pointed out earlier, although from a different vantage point.</p><ol><li>Automations need to be updated every time the underlying process changes or structure of information changes (which in startup land, is very, very often)</li><li>The customer charged with developing the automation needs to learn how to use the platform, and will likely need technical expertise to integrate a user database for instance (not a task for “Citizen Developers”)</li><li>Currently tooling is quite inflexible (especially when working with tabular data) and so the user is limited to performing the actions that the automation platform makes available. Anything bespoke has to developed outside the platform, stood up as an API, and connected to the automation platform, putting the customer back at square one.</li></ol><p>So for simple tasks, the existing suite of software can be a great fit, but as SMEs grow and mature, they are often motivated to roll out their own in-house solutions as they encounter more edge cases and their processes become more complex due to growing head-counts and increasingly cross-functional demands.</p><h3>What’s the next evolution of automation for the enterprise?</h3><p>Looking at existing solutions, we identified 3 key issues:</p><ul><li>Complex automations are too difficult to pull off without technical expertise</li><li>Automations in place are un-reliable in the face of small changes in the process</li><li>The steep learning curve hinders the the time required to see real value from the automations</li></ul><p>The team here at Solitude have collectively experienced the pain that comes with building and maintaining these low code automations, especially when the intended audience is non-technical. We wanted to make this experience so much better for customers and really deliver on that promise of letting you engage more complex, strategic work rather than get bogged down in the minutiae of the day to day.</p><p>We asked the question: what if we could completely re-invent the way we create and engage with automations from the ground up?</p><p>We drafted a short list for what this ideal solution would look like:</p><ul><li><strong>As simple as possible</strong>. As a customer, I should be able to construct an automation specific to my needs with a simple description. No faffing around with nodes, drop downs and complex diagrams or screen recordings.</li><li><strong>Zero hassle changes</strong>. As a customer, I shouldn’t have to worry about making changes to my process, my automations should change in response to my business, not the other way around.</li><li><strong>Immediate gains in productivity.</strong> Post setup, I shouldn’t need to spend weeks and months training a team or department on the platform. The onboarding process should be just as seamless on day 0 and day 10.</li></ul><p>Considering these factors, an emerging technology known as “Agents” came out on top as the most suitable approach to finally enable a refreshingly delightful user experience.</p><p>Agent technology is a form of generative AI built on-top of LLMs (large language models) similar to GPT-4. A key distinction here is that an Agent is <em>not</em> an LLM, it uses LLMs combined with advanced distributed systems architectures and prompt engineering techniques to do something really novel: take actions on behalf of the user, rather than just answering questions ChatGPT style.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gg073sVxuRi2srLvxOYlJg.png" /><figcaption>Amazing work by Lilian Weng if you want to learn more about agents at: <a href="https://lilianweng.github.io/posts/2023-06-23-agent/">https://lilianweng.github.io/posts/2023-06-23-agent/</a></figcaption></figure><p>What does this mean? As a user, you can connect your calendar to the agent and direct-message channels like whatsapp, slack and discord, and direct the agent to schedule meeting or follow-up calls with your customers whenever a it recognises positive signal from a customer interaction with your team on a chat channel or a new event is inserted into your CRM.</p><p>Better yet, you didn’t have to program any of that, or configure 100s of settings. As a customer, just connect your tools and tell the agent what do, just like you would a new hire. Of course what matters the most is that you remain in the driver’s seat for this interaction, so we’ve made it easy for you to ensure that you sign off on any action the agent proposes before it gets committed.</p><p>We have so much more we want to deep dive on, and we’ll certainly get into more details as we begin to release our weekly blog series. If what you’ve read here sounds interesting to you, we would love to have you join our waitlist to be invited to our early adopter’s program at <a href="https://solitude.ai/">https://solitude.ai</a></p><p>Sources:</p><ol><li>Why RPA fails to meet expectations: <a href="https://www.advsyscon.com/blog/why-rpa-fails-robotic-process-automation">https://www.advsyscon.com/blog/why-rpa-fails-robotic-process-automation</a></li><li>Automation doesn’t work if your enterprise doesn’t know what to automate: <a href="https://www.forbes.com/sites/forbestechcouncil/2022/03/16/automation-is-not-a-panacea-for-what-ails-the-enterprise/">https://www.forbes.com/sites/forbestechcouncil/2022/03/16/automation-is-not-a-panacea-for-what-ails-the-enterprise/</a></li><li>Agents as the next platform shift in enterprise software: <a href="https://www.mosaicventures.com/patterns/llm-agents-the-next-platform-shift-in-b2b-software">https://www.mosaicventures.com/patterns/llm-agents-the-next-platform-shift-in-b2b-software</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bcf9d79dca18" width="1" height="1" alt=""><hr><p><a href="https://medium.com/solitude-agents/is-the-automation-industry-stagnating-bcf9d79dca18">Is the automation industry stagnating?</a> was originally published in <a href="https://medium.com/solitude-agents">Solitude Agents</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>