<?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[The Lab @ Apply Digital - Medium]]></title>
        <description><![CDATA[The Lab is a small, dedicated, multidisciplinary team that explores and develops strategy for newer market-shifting technologies. Learn more: www.applydigital.com - Medium]]></description>
        <link>https://medium.com/the-lab-apply-digital?source=rss----f610836efb60---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>The Lab @ Apply Digital - Medium</title>
            <link>https://medium.com/the-lab-apply-digital?source=rss----f610836efb60---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 05:00:36 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/the-lab-apply-digital" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Leveraging AI to Save Time]]></title>
            <link>https://medium.com/the-lab-apply-digital/leveraging-ai-to-save-time-fcb4903f43aa?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/fcb4903f43aa</guid>
            <category><![CDATA[sports]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[efficiency]]></category>
            <category><![CDATA[time-management]]></category>
            <dc:creator><![CDATA[Sean Bowles]]></dc:creator>
            <pubDate>Thu, 24 Apr 2025 19:00:00 GMT</pubDate>
            <atom:updated>2025-04-24T20:26:43.481Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LXxGusJW47DQhLqVWvGoxQ.png" /></figure><p>We talk a lot about innovation in tech — but at the heart of innovation is time. Time to think, time to experiment, time to actually build the things that matter. That’s why, in The Lab at Apply Digital, we’ve focused much of our AI exploration on one core question: how do we give people their time back?</p><p>Whether we’re enabling internal teams or building Proofs of Concept (POCs) for our clients, our goal has been to develop AI-powered solutions that streamline workflows and make the day-to-day more efficient. Because let’s face it — when you’re spending 70% of your time in meetings reacting and talking about the work, and only 30% doing the work, there’s not a lot of room left for strategy or creativity.</p><p>That’s the challenge we set out to address through a series of AI POCs we recently presented on stage at the SBJ CAA World Congress of Sports — the largest and most prestigious sports business conference in North America. What follows is a look at two of those POCs, both designed to save time in very real, very tangible ways.</p><h3>The News Agent — Curated Intelligence on Demand</h3><p>Staying up to date shouldn’t feel like another full-time job. Whether you’re a sports executive, tech strategist, or someone who just wants to keep a finger on the pulse of your industry, the News Agent helps you stay informed without overwhelming you.</p><p>Built to deliver personalized daily briefings, the News Agent pulls from curated RSS feeds and other content sources to compile relevant updates, to you, into a single digest. Want to read it like a newsletter? Go for it. Prefer to listen to a podcast-style version on your commute? We’ve got that too.</p><p>It’s your own custom information assistant, cutting through the noise and making sure the stories that matter most are front and center — on your terms, and on your time.</p><h3>The Meeting Sidekick — Turning Conversations into Actions</h3><p>AI note-takers are helpful, sure — but we wanted to take that idea a step further. Enter the Meeting Sidekick, an agent designed not just to take notes, but to actually prepare you for meetings, and help you follow up with impact.</p><p>Here’s how it works: The Sidekick connects to your calendar and email, pulling together relevant attachments, past meeting notes, and context from ongoing threads. Ahead of your meeting, you receive a summary that gives you everything you need — without having to dig for it.</p><p>And afterward? The Sidekick can draft follow-up emails based on the conversation, saving you the mental load of summarizing next steps. It doesn’t replace your note-taker — it enhances it, operating independently and filling in the gaps with just the right level of context.</p><p>It’s like having an extra set of hands and a second brain — focused entirely on meeting efficiency.</p><h3>Let AI Help You Take Back Your Time</h3><p>These POCs are more than cool demos. They’re real examples of how applied AI can transform our day-to-day work — not in the future, but right now. By focusing on everyday friction points like keeping up to date, and meetings, we’ve shown that AI doesn’t need to be complex to be impactful. It just needs to be thoughtful.</p><p>And this is just the beginning.</p><p>Let’s build smarter — and give ourselves more time to do the work that matters most.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fcb4903f43aa" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/leveraging-ai-to-save-time-fcb4903f43aa">Leveraging AI to Save Time</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Enablement: How Apply Digital is Supercharging its Team]]></title>
            <link>https://medium.com/the-lab-apply-digital/ai-enablement-how-apply-digital-is-supercharging-its-team-d0e7baacac0e?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/d0e7baacac0e</guid>
            <category><![CDATA[ai-tools]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[development]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Sean Bowles]]></dc:creator>
            <pubDate>Fri, 31 Jan 2025 22:42:50 GMT</pubDate>
            <atom:updated>2025-02-05T00:16:09.275Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7_HFQrON64cDMTEkY4u_fg.jpeg" /><figcaption>Image generated using ChatGPT 4o based on the content of this post.</figcaption></figure><p>Here at Apply Digital, we’re always looking for ways to empower our employees and help them work smarter, not harder. One area we’ve been exploring over the past two years is leveraging the power of AI tools to supercharge teams and workflows across Apply. It’s been an interesting journey that we’ve decided to share with you in a series of posts delving into the work we’ve done and the decisions we’ve made along the way, starting with a quick recap of our journey thus far.</p><p>The purpose of The Lab here at Apply Digital is to experiment with nascent technologies both in the service of our clients and in the service of Apply Digital itself. While our early experiments with AI were more simplistic in nature, testing the capabilities of the technology in different ways, as we became more familiar with AI, we gradually moved towards more sophisticated solutions. For example, following the experiments we ran on <a href="https://medium.com/the-lab-apply-digital/state-of-intelligent-document-processing-idp-155338ebd4b9"><strong>Intelligent Document Processing</strong></a>, we built our own internal tool, which uses a blend of Large Language Models (LLMs) to analyze and extract key information from documents, allowing users to compare and contrast responses from four different models as well as optimize the responses for greater accuracy and comprehensiveness.</p><p>The next step in our journey with AI came in the form of developing custom GPTs to meet the specific needs of Apply’s various disciplines. For example, our EMEA Business Development team spends a lot of time responding to RFPs, as this is the common practice for the region. Working with them to build an RFP Engine that crafts responses to RFP questions by leveraging our extensive library of past submissions, they are now able to greatly reduce the amount of time spent on responding to an RFP. Similarly, working with the PMO team, we helped build and refine an SOW Engine that evaluates Statements of Work against pre-established rules and best practices, ensuring consistency and quality across all projects. These targeted AI-powered tools have proven incredibly valuable in optimizing workflows and freeing up our teams to focus on higher-level tasks, but as is the case with any internally developed tools, managing, improving and maintaining them comes with a cost.</p><p>With the demand for more AI tools increasing across the organization, we made it a priority for 2025 to choose a platform to host all of our custom GPTs and make them more accessible and, thus, more integrated into our daily work across Apply. Following a month-long trial, we decided to consolidate our efforts within Notion. With Notion AI, we are able to connect our AI tools seamlessly with Apply’s central knowledge base through integrations with Google Drive and Slack. In addition to supercharging our ability to access the vast knowledge within Apply, Notion AI provides us with a platform from which we can continue developing custom AI tools that are accessible to all Apply Digital employees at a reasonable cost.</p><p>Our journey into AI-powered workflows continues. We’re constantly exploring new tools and techniques to further empower Apply Digital and enhance our productivity as a whole. As we dive deeper into AI-enhanced workflows, we’ll continue to share our experiences and insights. Stay tuned for more updates on how we’re continuing to evolve and innovate in this exciting space!</p><p><em>This post was written with the assistance of Gemini Advanced. All images included in this post were produced using ChatGPT 4o.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d0e7baacac0e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/ai-enablement-how-apply-digital-is-supercharging-its-team-d0e7baacac0e">AI Enablement: How Apply Digital is Supercharging its Team</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Choosing the Right Code Assistant: A Comparative Analysis of Github Copilot, Gemini Code Assist…]]></title>
            <link>https://medium.com/the-lab-apply-digital/choosing-the-right-code-assistant-a-comparative-analysis-of-github-copilot-gemini-code-assist-6384e0132424?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/6384e0132424</guid>
            <category><![CDATA[github]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[gemini]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Meghna Lohia]]></dc:creator>
            <pubDate>Thu, 01 Aug 2024 19:26:37 GMT</pubDate>
            <atom:updated>2024-08-01T19:26:37.504Z</atom:updated>
            <content:encoded><![CDATA[<h3>Choosing the Right Code Assistant: A Comparative Analysis of Github Copilot, Gemini Code Assist, and Twinny + Codellama</h3><p>By Andrew Lee</p><p>In the fast-paced world of software development, code assistants are emerging as indispensable tools, boosting productivity and streamlining workflows. <strong>At Apply Digital, we’re committed to leveraging cutting-edge AI to empower our clients. In this spirit, we’ve conducted a rigorous evaluation of three leading code assistants: Github Copilot, Gemini Code Assist, and Twinny + Codellama. Our aim is to provide insights that help developers and businesses make informed decisions about which tool best suits their needs.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WY2trbpeX5Ig9hfD4gZdjA.png" /></figure><h3>Study Summary</h3><p>Our assessment highlighted Github Copilot as the best choice for mass adoption due to its consistently accurate responses and seamless user experience. Gemini, in particular, was more susceptible to hallucinations — a term referring to the generation of irrelevant or incorrect information by AI models. This issue is likely due to the relative newness of the Gemini and llama3 AI models compared to OpenAI’s Codex.</p><p>Twinny, although more cumbersome in user experience, operates on local hardware, addressing client security and confidentiality concerns. This makes it a suitable fallback option for more sensitive clients. Despite these differences, each tool has its strengths and areas for improvement:</p><h3>Methodology</h3><p>We conducted three distinct coding exercises to test the assistants:</p><ol><li><strong>Calculator Fill-in-the-Middle:</strong> Completing mathematical logic in partially-written classes.</li><li><strong>Password Cracker Functions:</strong> Generating functions for password validation, brute force attacks, dictionary attacks, and Markov chain attacks based on comment prompts.</li><li><strong>React Component Generation:</strong> Creating a batch of basic React components, from simple elements to a complete page with minor business logic.</li></ol><p>All code was generated solely by the code assistants, with no user input beyond initial notes.</p><h3>Test Results</h3><h4>Calculator Fill-in-the-Middle</h4><ul><li><strong>Github Copilot:</strong> Delivered excellent quality, understanding private scoping and providing a compositional approach without violating it. It had no unexpected behaviors and was easy to use.</li><li><strong>Gemini Code Assist:</strong> Good quality, with some challenges in accounting for private scoping and occasional unorthodox code suggestions. It experienced some hallucinations but was generally easy to use.</li><li><strong>Twinny/Codellama:</strong> Offered quality similar to Gemini, with slower response times due to local hardware limitations. It was moderately easy to use.</li></ul><h4>Password Cracker Functions</h4><ul><li><strong>Github Copilot:</strong> Provided good quality suggestions, with minor tweaks needed for the Markov chain function. It had no hallucinations and was easy to use.</li><li><strong>Gemini Code Assist:</strong> Exhibited less quality compared to Copilot in some functions, with logical fallacies in the brute force function and hallucinations in the Markov chain function. It was easy to use, with some helpful additional comments.</li><li><strong>Twinny/Codellama:</strong> Showed moderate quality, misunderstanding some requirements, and including logical fallacies. It was moderately easy to use, with some lag on complex functions.</li></ul><h4>React Component Generation</h4><ul><li><strong>Github Copilot:</strong> Delivered good quality, minimal tweaks required, and was able to incorporate previous suggestions. It had no hallucinations and was easy to use.</li><li><strong>Gemini Code Assist:</strong> Provided excellent quality with well-done component code and the use of React hooks. It added detailed comments and was easy to use.</li><li><strong>Twinny:</strong> Offered moderate quality, with some shallow suggestions and missing functionalities. It had unexpected looping behavior but was moderately easy to use.</li></ul><h3>Discussion</h3><p><strong>Github Copilot:</strong> While not flawless, Github Copilot provided mostly accurate suggestions. However, it requires users to have a deep understanding of their domain to prompt for optimal solutions effectively.</p><p><strong>Gemini Code Assist:</strong> Despite some challenges with hallucinations and weaker responses in certain areas, Gemini Code Assist showed strong potential under the right conditions, in particular with our component code generation test. With further development, it could become a formidable competitor.</p><p><strong>Twinny:</strong> Twinny delivered responses similar in quality to Gemini Code Assist, except in the components test. Running on local hardware makes it slower, but it offers significant security advantages, making it a viable option for sensitive projects.</p><h3>Conclusion</h3><p><strong>Key Takeaways:</strong></p><ul><li>Github Copilot currently stands out as the most reliable and comprehensive code assistant among the three tested.</li><li>Gemini Code Assist shows promise, particularly in specific tasks, and could become a major player with further development.</li><li>Twinny/Codellama offers a secure, albeit slower, alternative, making it a valuable option for sensitive projects.</li></ul><p>As AI continues to advance, Apply Digital remains dedicated to exploring and harnessing the latest innovations to deliver the most effective digital solutions for our clients. We’re particularly intrigued by the potential of leveraging our GPT-4o API through Twinny, as initial tests suggest this hybrid approach could offer an unparalleled combination of power, security, and efficiency. Stay tuned for further updates as we continue to push the boundaries of what’s possible in the realm of code assistance.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6384e0132424" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/choosing-the-right-code-assistant-a-comparative-analysis-of-github-copilot-gemini-code-assist-6384e0132424">Choosing the Right Code Assistant: A Comparative Analysis of Github Copilot, Gemini Code Assist…</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Revolutionizing Jira Tickets: How AI Can Boost Your Workflow]]></title>
            <link>https://medium.com/the-lab-apply-digital/revolutionizing-jira-tickets-how-ai-can-boost-your-workflow-96796880e4ee?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/96796880e4ee</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[jira]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Meghna Lohia]]></dc:creator>
            <pubDate>Wed, 12 Jun 2024 19:21:45 GMT</pubDate>
            <atom:updated>2024-06-12T19:21:44.923Z</atom:updated>
            <content:encoded><![CDATA[<p>By Joe Hankin</p><p>While crafting well-defined Jira tickets is essential for a smooth development and effective product management, the task itself can feel like a chore. It eats into valuable time that could be spent on creative problem-solving and innovation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8eDKCl6zY5Qh2VxvZGiQFg.jpeg" /></figure><p>Jira tickets aren’t just a to-do list; they’re a blueprint for successful development. Best practices dictate that tickets follow a specific formula: clearly defined outcomes, a concise user story outlining the purpose, and unambiguous acceptance criteria for success. Adhering to this structure ensures that every ticket has the essential information developers need for timely and accurate execution.</p><p>Unfortunately, even with this understanding, manually writing these detailed tickets can be a bottleneck in the development process. That’s why we’ve harnessed the power of AI to do just that.</p><p><strong>Our AI-Powered Jira Ticket Assistant</strong></p><p>We’ve leveraged OpenAI’s ChatGPT to create a Jira Ticket assistant — a game-changer for ticket creation. Our assistant transforms any request into well structured, detailed tickets.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WvHnxJuy6JW1gs8Cu8Psiw.png" /></figure><p>Our ticket assistant boasts two key advantages: speed and precision. By automating ticket creation, you’ll experience a significant reduction in time spent on this repetitive task. <strong>Estimates suggest a 2x improvement in ticket creation speed</strong> — a valuable boost for any team. But speed isn’t all that matters. Our assistant excels at capturing even the most complex feature requests with exceptional accuracy. This translates to less confusion, smoother communication, and a team consistently aligned on project goals.</p><p>We’re committed to using AI for real-world business benefits, simplifying workflows so your team can focus on delivering exceptional results.</p><p><strong>Ready to see AI transform your Jira workflow? Contact Apply Digital today to learn more.</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=96796880e4ee" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/revolutionizing-jira-tickets-how-ai-can-boost-your-workflow-96796880e4ee">Revolutionizing Jira Tickets: How AI Can Boost Your Workflow</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How ChatGPT Can Spark Animation Creativity and Improve the Customer Experience]]></title>
            <link>https://medium.com/the-lab-apply-digital/how-chatgpt-can-spark-animation-creativity-and-improve-the-customer-experience-a778a4efa64a?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/a778a4efa64a</guid>
            <category><![CDATA[motion-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Meghna Lohia]]></dc:creator>
            <pubDate>Tue, 30 Apr 2024 17:52:30 GMT</pubDate>
            <atom:updated>2024-05-15T22:20:31.366Z</atom:updated>
            <content:encoded><![CDATA[<p>Rico Boas, Shafakat Reshamvala, Mitsu Yajima</p><p>Animation is a key component of captivating and engaging users. It’s not just about eye-catching designs; it’s about creating intuitive, user-friendly experiences. While once a novelty, animation is now an essential part of digital design.</p><p>Crafting high-quality animations is both an art and a science. Mastering this process requires a high level of technical expertise. The more complex the design, the more time it takes to bring it to life. Larger file sizes also come with the risk of slower page load times, diminishing user satisfaction. Consider this: a four-second delay in page loading can drive away <a href="https://techreport.com/statistics/website-load-time-statistics-data/">40% of your audience</a>.</p><p>So, how can teams accelerate the design workflow, unlock creativity, and create the best user experience?</p><p>At Apply Digital, we’re redefining the animation process through innovative AI applications. This post delves into our unique approach to creating captivating micro-interactions that delight and surprise our users.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/853/1*Oh_OXVKycqzJD-DtXOS5Ng.gif" /></figure><p><strong>The Animation Process</strong></p><p>Designing lifelike animation is a complex and time-consuming process. Utilizing tools like After Effects, animators have to set keyframes and animate elements frame by frame. This process has specific constraints, such as efficiency, creative limitations, and accuracy.</p><p>These factors are important to consider because animations need to accurately mimic the laws of physics. Recreating natural movements, whether it’s the ripple of water or the simple bounce of a ball, requires creativity and a deep understanding of physics. While expertise can guide designers to use the correct mathematical expressions and produce realistic animations, it’s not always a straightforward process. For instance, designers need to know the right equation <em>and </em>how to apply it intuitively. The challenge is to make this skill level more accessible to a broader range of designers.</p><p>Depending on the complexity, it can take around 2 to 6 weeks to design and animate just one minute of motion graphics. Given this timeline, there’s an apparent opportunity to innovate and accelerate the process.</p><p>Enter artificial intelligence (AI) and ChatGPT. In particular, AI and language learning models (LLMs) have opened new avenues for approaching complex animations. These tools allow us to translate artistic concepts into efficient, accurate, and adaptable animations without compromising the creative essence.</p><p>At the core of this transformation is an After Effects feature known as Expression. In short, it allows designers to create animation via equations that dictate motion. Using ChatGPT alongside this feature, our team is moving towards a coding and math-based approach to animation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/590/1*BQNe8mprdROBxcNizKJocQ.gif" /></figure><p><strong>The Impact</strong></p><ul><li><strong>Accessibility:</strong> ChatGPT demystifies Expression, making it accessible to designers who are less experienced with equations. This expands the creative capabilities of our entire design team.</li><li><strong>Creativity and Adaptability:</strong> Designers can rapidly prototype and adjust animations in response to evolving project requirements. This agility encourages taking more creative risks, recognizing adjustments can be made swiftly and efficiently.</li><li><strong>Efficiency:</strong> The AI-generated code we’ve created is smaller and more efficient, resulting in smoother and lighter animations — a reduction from 3MB to a mere 150KB. This also means faster loading times, which is particularly important for mobile experiences.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*8c3gKjJAPGP_nzfppW0dhw.gif" /></figure><p>It’s important to note that ChatGPT and the integration into After Effects aren’t perfect. They’re not a replacement for a designer’s creative expertise and there is still a need for expert motion refinement, especially because these tools are constantly evolving.</p><p><strong>Conclusion</strong></p><p>Our team has explored how ChatGPT can create efficiencies in animation production, spark creativity, and enhance the user experience through faster load times and smooth animations. As we continue to embrace AI in our workflows, stay tuned for future updates on how teams across Apply Digital are leveraging this transformative technology.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a778a4efa64a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/how-chatgpt-can-spark-animation-creativity-and-improve-the-customer-experience-a778a4efa64a">How ChatGPT Can Spark Animation Creativity and Improve the Customer Experience</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How we’re using AI to Revolutionize User Segmentation]]></title>
            <link>https://medium.com/the-lab-apply-digital/how-were-using-ai-to-revolutionize-user-segmentation-e0034a2e76a7?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/e0034a2e76a7</guid>
            <dc:creator><![CDATA[Meghna Lohia]]></dc:creator>
            <pubDate>Thu, 18 Apr 2024 22:06:31 GMT</pubDate>
            <atom:updated>2024-05-15T22:23:16.716Z</atom:updated>
            <content:encoded><![CDATA[<p>Calvin Chan and Evan Situ</p><p>User segmentation is essential for enhancing the user experience. Personalized calls to action, for instance, can lead to a <a href="https://www.demandsage.com/personalization-statistics/">202% increase in conversions.</a> At Apply Digital, we’re enhancing user segmentation with cutting-edge AI and cloud tools. This approach unlocks deeply personalized experiences that traditional methods simply can’t achieve.</p><p><strong>Our AI-Driven Approach</strong></p><p>Traditional user segmentation models have their limits. They rely on rigid rules and complex structures, demanding constant updates to keep up with changing user behavior.</p><p>Our approach uses OpenAI and NLP to transform broad cohort-based segments into hyper personalized individual profiles. This approach bypasses the expensive, pre-packaged user segmentation tools in the market. Our model is backed by an architecture that treats user data as dynamic input, feeding into AI systems for real-time segmentation. This architecture enables us to build our own segmentation model, offering the power of 3rd party tools without the premium price. Our results not only enrich our understanding of our users but are also interpretable and actionable.</p><p><strong>Architectural Insights</strong></p><p>Our exploration of AI segmentation takes two forms. Language-based segmentation and action-based segmentation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YTkWlwHJ0ozn_jyxb9AXiw.png" /><figcaption>Language-Based User Segmentation</figcaption></figure><p>In language segmentation, we use large volumes of chat data to extract specific user attributes. To achieve individual segmentation, our AI model demands extensive user data. We accomplished this by developing chat products that deliver genuine value and utility to end users and seamlessly integrating analytics into these chat experiences and related web properties.</p><p>We process chat logs via Cloud Scheduler. ChatGPT and Cloud Functions then extract and map key user attributes. Our model goes beyond surface data, capturing twice as many consumer attributes with a 90% accuracy rate in mapping.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bPlcUD0nBtgvWhuZzi9GFQ.png" /><figcaption>Action-Based User Segmentation</figcaption></figure><p>For action-based segmentation, we analyze user interactions with our website. OpenAI uses these behavior details for rapid and accurate user categorization. Dataflow then processes website interactions, creating a rich picture of user behavior that OpenAI leverages for predictive segmentation. This method has accelerated our model training and expanded our understanding of users.</p><p>For example, we provided OpenAI with real-time data on our Apply Digital website, including website navigation patterns, time on pages, and visit frequency. We defined segment categories like “Potential Clients,” “Industry Professionals,” or “Job Seekers.” This data stream enriches our consumer profiles within our CDP daily.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FvtmqEZO8jxt4ZZyOu4uXQ.png" /></figure><p><strong>Conclusion</strong></p><p>Our shift towards AI-driven, individualized segmentation dramatically enhances personalization and enriches user profiles. With our AI-driven approach, we were able to double the number of captured user attributes, and this robust model was built and implemented in mere weeks at a fraction of the cost of third-party tools. Our approach demonstrates the power of AI to transform data into actionable insights, enhancing the user experience.</p><p>Curious how AI can revolutionize your processes? Let’s explore solutions together — contact Apply Digital today.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e0034a2e76a7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/how-were-using-ai-to-revolutionize-user-segmentation-e0034a2e76a7">How we’re using AI to Revolutionize User Segmentation</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Integrating Artistic QR Codes Using AI]]></title>
            <link>https://medium.com/the-lab-apply-digital/integrating-artistic-qr-codes-using-ai-ff6c6a3b9511?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/ff6c6a3b9511</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[ai-image-generator]]></category>
            <category><![CDATA[qr-code]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Gustavo Salas]]></dc:creator>
            <pubDate>Wed, 24 Jan 2024 01:22:23 GMT</pubDate>
            <atom:updated>2024-01-24T01:22:23.884Z</atom:updated>
            <content:encoded><![CDATA[<p>QR codes are the modern bridge to streamline communication between the physical world and the digital one. However, they still look like an afterthought in the corner of a graphic piece, disrupting the intended reading flow of an image (and making our graphic designers and artists cry).</p><p>Our objective was to leverage AI to produce a QR code that was beautiful as a standalone piece while maintaining its readability, becoming part of the message where it is embedded. We initially explored existing products that could help us create artistic QR Codes using AI.</p><p>AI tools have been evolving at breakneck speeds, so there must be tools that could provide what we envisioned, but after testing more than ten different AI-powered QR code generator software, most produced either generic results or unreadable QR codes. Only one tool generated decent outcomes, but only when specific styles were used with our prompts, like fantasy, anime, or robots.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZGwMsSWwIfSZaWfWtMzT8Q.png" /><figcaption>Generated in QR Code AI</figcaption></figure><p>After this ‘walk in the desert’ experience, we began experimenting with Stable Diffusion online via Run Diffusion, which improved results but left the QR code somewhat separate from the visual output. While it blended our vision better, achieving the desired outcome required further exploration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/0*jICrgiOA3OyMLNHC" /><figcaption>Local Stable Diffusion UI</figcaption></figure><p>As working with Run Diffusion became more challenging — due to the original checkpoint tested being removed — the settings we’d been testing no longer provided the results we were after, so we decided to take Stable Diffusion local. However, running locally required immense computing power and was very slow. Setting it up required downloading several checkpoints, models, and preprocesses. The whole process was technically demanding and painfully slow.</p><p>As we were trying to optimize the procedure, we found another online tool called “Think Diffusion”, that uses Stable Diffusion to create custom outputs with a similar but more user-friendly interface than Run Diffusion. This online tool also featured a pricing model reminiscent of old-school internet cafes, charging for the time you use the app (versus the others, which all charged per image). Think Diffusion broke its process down into two workflows:</p><ol><li><strong>Text-to-Image: </strong>Generates an image alongside the QR code, merging the two.</li><li><strong>Image-to-Image: </strong>Blends an existing image with the QR code via ControlNet, a neural network structure that allows us to fine-tune (further testing is required as this round provided results that either looked good, but the QR code wasn’t scannable or they were scannable but looked odd).</li></ol><p>So where did we end up? In certain circumstances, you can quickly and easily generate custom QR codes as long as your desired output is within a small list of preset themes. For the majority of use cases, it’s possible, but it requires time to experiment and fine-tune to create consistent, high-quality outputs. Along the way, we’ve learned that specific scenarios won’t work; for example, images without high contrast cannot produce a scannable code. But even a perfect starting image needs help to become a great-looking QR code.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2Oj8U-p6BO9nbpADuv_Qnw.png" /><figcaption>Left was generated with Stable Diffusion and the right was edited so it was scannable.</figcaption></figure><p>The quest for the ideal technique to use generative AI in producing artistic QR Codes on demand continues. If you’re interested in a tool like this for your brand, we’re ready to help craft one for your specific needs.</p><p><em>Co-written by </em><a href="https://www.linkedin.com/in/chloekeogan/">Chloë Keogan-Grant</a>, <a href="https://www.linkedin.com/in/wellsstringham/">Wells Stringham</a> and <a href="https://www.linkedin.com/in/andresvillaseca/"><strong><em>A</em></strong>ndres Villaseca</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ff6c6a3b9511" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/integrating-artistic-qr-codes-using-ai-ff6c6a3b9511">Integrating Artistic QR Codes Using AI</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[State of Intelligent Document Processing (IDP)]]></title>
            <link>https://medium.com/the-lab-apply-digital/state-of-intelligent-document-processing-idp-155338ebd4b9?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/155338ebd4b9</guid>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[generative-ai]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[openai]]></category>
            <dc:creator><![CDATA[Tyler Nee]]></dc:creator>
            <pubDate>Wed, 05 Apr 2023 16:21:46 GMT</pubDate>
            <atom:updated>2023-04-05T16:21:34.242Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L6nwJPV29EbrpFlncgfzqQ.png" /></figure><p>Say goodbye to manual data entry and hours of tedious work reading documents! Intelligent Document Processing (IDP) is the future of streamlining the analysis, extraction and interaction of data from documents. By leveraging cutting-edge AI models such as GPT-4, IDP automates document processing by enabling people to speak and ask questions to a document as if it were a person. This saves businesses time and money while dramatically reducing human errors. In this article, we explore the state of IDP, the best-in-class tools available, and how companies can integrate IDP into their workflows to unlock its full potential. If you’re ready to supercharge your business operations and productivity, keep reading to discover the power of IDP.</p><p><strong>The IDP Landscape</strong></p><p>The current landscape of Intelligent Document Processing (IDP) is rapidly evolving alongside the rapid advancement in AI technology. The IDP market is expected to grow at a <a href="https://www.globenewswire.com/news-release/2023/02/28/2617349/0/en/Intelligent-Document-Processing-Market-Size-Share-Trends-Analysis-Report-By-Component-By-Technology-By-Deployment-By-Organization-Size-By-End-use-By-Region-And-Segment-Forecasts-20.html#:~:text=Intelligent%20Document%20Processing%20Market%20Growth%20%26%20Trends&amp;text=The%20market%20is%20expected%20to,requiring%20efficient%20document%20processing%20solutions.">CAGR of 29.7%</a> between 2022 and 2030, reaching a staggering <a href="https://finance.yahoo.com/news/intelligent-document-processing-market-worth-114500265.html#:~:text=SAN%20FRANCISCO%2C%20Jan.,29.7%25%20from%202022%20to%202030.">$11.6 billion</a> by 2030. With the exponential growth of unstructured data and the rising need for efficient data processing, IDP is quickly becoming a vital tool for businesses across a range of industries.</p><p>The key players in the IDP market are investing heavily in research and development to create increasingly sophisticated solutions that can not only answer complex questions, but also handle multi-document processing and data visualizations. Some early market leaders include Explainpaper, ChatDOC, and Chatbase who are driving innovation in different avenues within IDP which we’ll dive into shortly.</p><p>From finance and healthcare to sports and entertainment, IDP is revolutionizing the way businesses handle document processing, leading to increased productivity, improved insights gathering, and operational efficiency. The future of IDP is bright, and this technology will play a vital role in shaping how companies operate in the next few quarters, let alone years.</p><p><strong>Best In-Class IDP Tools</strong></p><p>In the world of Intelligent Document Processing (IDP), the competition is fierce, and the best-in-class tools are constantly evolving to meet the ever-growing demands of businesses.</p><p>One such tool is <a href="https://www.explainpaper.com/">Explainpaper</a>, which is revolutionizing the way people understand complex topics. With Explainpaper, users can quickly gain a deep understanding of even the most complicated concepts, providing detailed answers to help simplify even the most challenging of topics. By using Explainpaper, businesses can save time and resources while increasing efficiency and accuracy in their document processing tasks.</p><p>Another best-in-class IDP tool that is taking the market by storm is <a href="https://chatdoc.com/">ChatDOC</a>. ChatDOC is a tool that formulates contextualized summaries of large PDF files with incredibly accurate cited sources. With ChatDOC, people can easily reference where the information was gathered from, helping to ensure the accuracy and integrity of the information. This tool is a game-changer for businesses that need to quickly extract information from large documents, saving time and resources while increasing accuracy and efficiency.</p><p>To finish off the list, <a href="https://www.chatbase.co/">Chatbase</a> is a best-in-class IDP tool that is changing the game for businesses. Chatbase can convert any document into a chatbot that answers simple questions on a particular file, allowing businesses to quickly provide answers to their customers without the need for manual intervention. This tool can be exported into various platforms such as an internal Slack channel, a public-facing website, or even within a mobile app. With Chatbase, businesses can save time and resources while providing exceptional customer service, leading to increased customer satisfaction and loyalty.</p><p><strong>Integration of IDP into Workflows</strong></p><p>Integrating Intelligent Document Processing (IDP) into workflows is changing the game for businesses across a range of industries. In finance, IDP is transforming the way businesses analyze financial statements and charts. With IDP tools, finance professionals can easily extract data from PDFs of annual reports on publicly listed companies like Nike or Lululemon, freeing up valuable time and resources to focus on more complex financial questions such as valuation.</p><p>The sports and entertainment industry is also leveraging IDP to streamline their document processing tasks. With IDP tools like Chatbase, businesses can convert contracts, agreements, and other legal documents into a chatbot that answers simple questions, allowing teams to quickly access vital information without the need for manual intervention. This technology is revolutionizing the way sports and entertainment businesses operate, improving their efficiency and allowing them to focus on providing the best possible experience for their fans and customers.</p><p>Consumer packaged goods (CPG) companies are also turning to IDP to optimize their workflows. With IDP tools like Explainpaper, CPG companies can quickly gain a deep understanding of consumer trends and preferences, allowing them to create better products that meet the needs of their customers. By using IDP to automate data processing tasks, CPG companies can save time and resources while improving accuracy, leading to increased customer satisfaction and loyalty.</p><p><strong>Recommendations</strong></p><p>Choosing the right Intelligent Document Processing (IDP) tool can be as tricky as navigating a corn maze blindfolded. With so many options available, it’s essential to consider several factors before selecting a tool. One crucial factor to consider is the tool’s learning curve. While some IDP tools may require a Ph.D. in computer science to use effectively, others are as easy as drag &amp; drop. It’s essential to choose a tool that fits your specific needs.</p><p>Data privacy compliance and security is more important than ever. In light of the recent hiccups with OpenAI leaking ChatGPT conversations, choosing an IDP tool with robust security should be a top priority. The last thing you want is expose your company to cybersecurity risks because confidential business data got leaked to the wrong hands.</p><p>When it comes to IDP tools, it’s all about the return on investment. You want a tool that saves you time and resources while improving the accuracy of your document processing tasks. If you’re feeling frustrated, it’s probably the wrong tool.</p><p>Last but not least, it’s important to choose a tool that’s not too green. While it’s exciting to be on the cutting edge of technology, early-stage IDP tools may not have all the features and functionality you need. You don’t want to put all your eggs in one basket and end up with a tool that shuts down or can’t meet your specific needs.</p><p><strong>Conclusion</strong></p><p>In conclusion, Intelligent Document Processing (IDP) is a game-changer for businesses looking to improve their document processing workflows. With best-in-class tools like Explainpaper, ChatDOC, and Chatbase, businesses can achieve unprecedented levels of efficiency and accuracy in their document processing tasks. However, it’s essential to choose the right IDP tool based on factors like the learning curve, data privacy compliance, security, ROI, and tool maturity.</p><p>Looking forward, the future of IDP is bright. With advancements in generative AI, LLMs and machine learning, we can expect to see IDP tools becoming even more sophisticated and powerful with multi-document processing and possibly even audio-based conversations. As IDP tools continue to evolve, they will become even more accessible and user-friendly, helping businesses of all sizes achieve significant productivity gains. Overall, businesses that adopt IDP tools today will be well-positioned to thrive in this new age of AI.</p><p><em>Apply Digital solves complex problems with well-executed digital solutions for some of the most respected brands in the world. We help our clients gain a competitive advantage and delight users with expertly executed digital solutions. Learn more about our services at </em><a href="http://www.applydigital.com/"><em>www.applydigital.com</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=155338ebd4b9" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/state-of-intelligent-document-processing-idp-155338ebd4b9">State of Intelligent Document Processing (IDP)</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Starbucks’ Odyssey: The Future of Loyalty Or Is It Just Hype?]]></title>
            <link>https://medium.com/the-lab-apply-digital/starbucks-odyssey-the-future-of-loyalty-or-is-it-just-hype-c8b2ad0e7b62?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/c8b2ad0e7b62</guid>
            <category><![CDATA[web3]]></category>
            <category><![CDATA[loyalty]]></category>
            <category><![CDATA[strategy]]></category>
            <category><![CDATA[starbucks]]></category>
            <category><![CDATA[nft]]></category>
            <dc:creator><![CDATA[Tyler Nee]]></dc:creator>
            <pubDate>Mon, 06 Mar 2023 23:57:09 GMT</pubDate>
            <atom:updated>2023-03-06T23:48:17.098Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hTDyZHSGKSGJkDgcrzGShw.png" /><figcaption>Starbucks’ Latest Digital Stamp Collection</figcaption></figure><p>Since the beta officially launched on Dec. 8, Starbucks continues to onboard new customers by the day. Now that Odyssey has been out for three months, we’re breaking down whether the program is truly inventing the future of loyalty or if it’s just hype.</p><p>To summarize, Starbucks Odyssey is a first-of-its-kind Web 3.0 loyalty program that launched its Beta program in December 2022. How it works is customers log in with their Starbucks Rewards account and participate in challenges to earn “NFTs” called Journey Stamps. Through completing challenges and earning stamps, customers earn points that unlock unique Starbucks rewards, such as virtual martini-making classes or unique access to artist merchandise. For a full strategic analysis of Starbucks Odyssey, you can read more about it <a href="https://medium.com/@tyler.nee/strategic-analysis-starbucks-odyssey-web-3-0-nft-based-loyalty-program-3d16c5762267">here</a>.</p><h3><strong>The Experience So Far</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GAapLK-a4qokDlM5qn_4Vw.png" /></figure><p><strong>Single Sign-On</strong></p><p>Starbucks Rewards has over 31M members, making it one of the most successful loyalty programs on the planet. Thus, Starbucks doesn’t need to start from scratch. So when designing Odyssey, they’ve decided to use Single Sign-On to leverage the community of already engaged customers from Starbucks Rewards. Customers can log in to the Odyssey Web app using their existing Starbucks Rewards login credentials when they get their invitation to join the beta. This provides a seamless experience so customers don’t need to create a separate account and can dive right into the experience. In addition, this allows Starbucks to gather more data about these customers to better understand their preferences and behaviors and provide more personalized offerings. Along with linking their Rewards and Odyssey accounts, Starbucks also manages customers’ digital wallets within Odyssey in partnership with Nifty Gateway. So customers don’t need to worry about where their digital collectibles will be stored as it is all embedded into the Odyssey app.</p><p><strong>Potential To Be Big</strong></p><p>While over 1,000 lucky customers received exclusive access to Odyssey’s beta program, their new Web 3.0 loyalty program is still in the experimental phase. Starbucks hasn’t released official engagement metrics, but we can get a sense of the levels of engagement by scanning Nifty Gateway.</p><p>At the time of writing, Starbucks has a total transaction volume of $214K on its digital collectibles. They receive a 10% royalty for each transaction on Nifty Gateway, which means they’ve earned $21.4K in revenue. The prices of each collectible range from $80 to $2,000.</p><p>In addition, they’ve released four journey stamps to date, which means that there have been four challenges that customers have been able to participate in.</p><p>Here’s a list of the current stamps:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sYgNXlA5eZ3d8Vl6j8W12g.png" /></figure><p>If we look closer, less than 15% of customers are selling/trading their digital collectibles on average. This means that a majority of customers are holding onto the collectibles. They’re probably hoping their collectibles will increase in value or there will be future hidden perks for owning a given collectible.</p><p>Starbucks incentivizes consuming loyalty points by attaching an expiry date to ones associated to digital collectibles. But the points, which you can redeem for in-store drinks or merchandise, seem to be a minor perk compared to the speculative value of stamps on the NFT marketplace. A Holiday Cheer Edition 1 most recently sold, as of writing, for $1,700.</p><p>If Starbucks can onboard 5M customers to Odyssey, and these numbers hold, they’re looking at an approximate transaction volume of $705M, meaning $70.5M in revenue for them in year 1. So there’s potential for this to be a significant revenue stream for Starbucks.</p><p>Despite this potential upside, Starbucks has been careful about rolling out a Web 3.0-based product in a way that hinges on customer experience. This careful approach that balances adoption, technology viability and risk assessment, and a sensitivity of public sentiment, is one that we and other brands can learn from.</p><h3><strong>Positive Customer Sentiment</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/580/1*JSm9oQ-40AAc8KyReKr08w.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/590/1*cRFb5ndvshwpLXOeYJLchA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/590/1*MMMVphZQ8IhMYPNf4zE1oQ.png" /></figure><p>Overall, the customer sentiment has been pretty positive. Many customers like the ability to “monetize” their time and efforts by participating in challenges. They enjoy being incentivized to partake in actions they usually would already do, such as buying coffee or giving friends gift cards. Although, as already mentioned, there have been low volumes of transactions.</p><p>Since Starbucks has such a strong customer base, Odyssey provides them with a new way to interact with the brand outside of when they’re purchasing coffee. The Web 3.0 Loyalty program provides a gamified experience that brings a deeper connection with consumers that spans both physical and digital. As brands look to the future of loyalty, programs like Odyssey could be early examples of how companies evolve their engagement strategies with customers, especially as we approach the rise of the Metaverse.</p><h3><strong>But Odyssey is not Web 3.0</strong></h3><p>Starbucks has made a big deal about Odyssey being a Web 3.0 loyalty program, but is it?</p><p>They’ve removed Web 3.0 jargon such as NFTs, tokens, and crypto-wallets from any customer-facing marketing copy and within the Odyssey experience. They have been abstracting the association to Web 3.0, presumably to emphasize the experience it enables.</p><p>Decentralized peer-to-peer transactions–touted as one of blockchain’s value propositions–is disincentivized in the Odyssey program. When you take your digital collectibles outside the Starbucks ecosystem, you lose your loyalty points and Starbucks based perks. Doing so decouples it from your Starbucks Rewards account and it then only lives within the Nifty Gateway marketplace.</p><p>In addition, if you look under the hood of Nifty Gateway, the majority of transactions for digital collectibles within the platform, including Starbucks’, may not actually be recorded on the blockchain. Nifty’s custodial platform stores collectibles in the Nifty Gateway Omnibus wallet and transacts within that wallet–not on the blockchain–to sidestep fees associated with blockchain transactions. Outside of these use cases, Nifty Gateway treats digital assets like any other before Web 3.0. This leaves us to question if any of Starbucks’ digital collectibles are actually recorded on the blockchain at all.</p><p>Starbucks has abstracted away so many of the components that would justify Odyssey as a Web 3.0 application or a decentralized platform. There are facets within the stack that allows for decentralization, but for the most part, it’s not. This seems more like a bridge to Web 3.0 or rather just hype.</p><p>But Starbucks has definitely considered the tradeoffs.</p><h3><strong>Risks with Web 3.0 Adoption</strong></h3><p>Firstly, Web 3.0 lacks centralized control. This means that brands operating in this space may be subject to market fluctuations and regulations that could significantly impact their brand image. Companies would be exposed to hacking, data breaches, and other security issues without centralized control. Not to mention, brands would need to embrace the freedom that users have in Web 3.0 by relinquishing control of user behaviour and designing experiences that hedge against bad actors.</p><p>Secondly, investment in development costs for Web 3.0 may not be justified for all brands. While the benefits of a decentralized web are tempting, it may not make financial sense for brands with more straightforward products or services to make significant investments in this area. In addition, building a presence on a new platform in a new ecosystem can be challenging and expensive, resulting in long-term financial strain for a brand.</p><p>In addition, a heavy focus on reducing user friction is a major challenge for brands operating in the Web 3.0 space. To attract and retain customers, companies must design user experiences that are seamless and intuitive. Given the complexity of Web 3.0 technologies, brands may struggle to achieve this, resulting in customers losing interest or abandoning the platform altogether.</p><p>Moreover, there are uncertainties around taking assets outside the ecosystem. While blockchain technology provides a high degree of transparency and immutability, taking assets outside the ecosystem may come with risks. For example, the transfer of digital assets may incur high transaction fees, and there is no guarantee that these assets will retain their value in the long term.</p><p>Finally, while Web 3.0 loyalty platforms could bring switching costs to zero, this could potentially result in brand churn. Brands may have to work harder to retain customers, as they can move between loyalty platforms at no cost. This can create significant challenges for brands to build loyalty and maintain their customer base.</p><p><strong>Conclusion</strong></p><p>Starbucks Odyssey has the potential to grow into a key part of Starbucks’ loyalty offering, engaging customers outside the bounds of buying coffee. Their approach sets them up early to understand the risks and value of experiences based on new Web 3.0 tools. Even though we call out that Odyssey only scratches the surface of Web 3.0 technology — they are experimenting with the ideas of collecting NFTs and peer-to-peer experiences — new behaviors that don’t currently exist in the mainstream. Starbucks is cleverly balancing Web 3.0 as a marketing tool, while abstracting it from the actual experience. This does foreshadow how Web 3.0 tools may be adopted in the future — completely invisible to the user. Starbucks is paving a path for how a global brand might leverage Web 3.0 for consumer loyalty — other brands would benefit from following this case study as they continue to explore, test, and gain experience with these new tools, behaviors, and ideals.</p><p><em>Apply Digital solves complex problems with well-executed digital solutions for some of the most respected brands in the world. We help our clients gain a competitive advantage and delight users with expertly executed digital solutions. Learn more about our services at </em><a href="http://www.applydigital.com/"><em>www.applydigital.com</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c8b2ad0e7b62" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/starbucks-odyssey-the-future-of-loyalty-or-is-it-just-hype-c8b2ad0e7b62">Starbucks’ Odyssey: The Future of Loyalty Or Is It Just Hype?</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Intro to Generative AI — Images (Stable Diffusion, Midjourney, Dall-E)]]></title>
            <link>https://medium.com/the-lab-apply-digital/intro-to-generative-ai-images-stable-diffusion-midjourney-dall-e-7812af8e8a98?source=rss----f610836efb60---4</link>
            <guid isPermaLink="false">https://medium.com/p/7812af8e8a98</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[content-creation]]></category>
            <category><![CDATA[generative-art]]></category>
            <category><![CDATA[stable-diffusion]]></category>
            <category><![CDATA[midjourney]]></category>
            <dc:creator><![CDATA[Tyler Nee]]></dc:creator>
            <pubDate>Thu, 01 Dec 2022 00:30:47 GMT</pubDate>
            <atom:updated>2022-12-01T00:31:32.786Z</atom:updated>
            <content:encoded><![CDATA[<h3>Intro to Generative AI — Images (Stable Diffusion, Midjourney, DALL-E)</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C67xeBQQF0GW_yol4fACxg.png" /></figure><p>Generative AI is one of the current zeitgeists for disruptive innovation. Due to recent advancements in AI research and development, AI-generated content has been flooding social feeds. Thus, we’re experiencing AI summer. Contrary to the concept of crypto winter, AI summer refers to a period of increased interest, development, and funding for the space. In short, Generative AI is the application of machine learning models that produce predictive outputs based on inputs such as written text prompts, images, and video.</p><p>There are currently <strong>five main branches of Generative AI:</strong> images, audio, text, code, and video. In this article, we’ll focus on images, the leading players in this subcategory, and what the next 12–18 months could look like for companies looking to navigate this new technology.</p><h3><strong>Here are some examples of Generative AI:</strong></h3><p>The following images have been created using Open AI’s DALL-E Text-to-Image Model. You can try it out for yourself <a href="https://openai.com/blog/dall-e-now-available-without-waitlist/">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*3vRjU5mHU53Frdd0" /><figcaption>If I’m a cartoon artist and need inspiration, I can write a prompt such as “cartoon dogs floating in space” to generate this image.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Byw-KbSkelefD2va" /><figcaption>If I’m a travel blogger and need stock images to add to my articles, I can write a prompt such as “wind turbines located at the top of a cliff overlooking a vast ocean” to generate this image.</figcaption></figure><p>This is a remarkable evolution in AI technology, and we have yet to unlock its full potential. So, where did this all begin?</p><h3><strong>The Timeline</strong></h3><p>In 2017, Google Brain originally started working on predictive text generation via Large Language Models (LLMs) to help with the problem of auto-translation of languages while preserving context. Slowly, other companies such as Open AI began to adopt and build upon this work leading to breakthroughs such as Open AI’s natural language generator, GPT-3, released in Jun. 2020, and their first text-to-image generator, DALL-E, in Jan. 2021. The latest advancement in generative AI has been the development of <a href="https://arxiv.org/abs/2112.10752">latent diffusion models</a> in Dec. 2021. This discovery has helped companies generate higher-resolution images which has led us to the current AI gold rush. Fast forward to the summer of 2022, DALL-E2, Midjourney, and Stable Diffusion were released, and the tech world was taken by storm.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*51Q1ip2JVrXuWsVu" /><figcaption>Src: <a href="https://octoml.ai/blog/from-gans-to-stable-diffusion-the-history-hype-and-promise-of-generative-ai/">https://octoml.ai/blog/from-gans-to-stable-diffusion-the-history-hype-and-promise-of-generative-ai</a></figcaption></figure><p>Now that we understand a bit of the history, let’s dive into the leading players within Generative AI images: Stable Diffusion, Midjourney, and DALL-E.</p><h3><strong>Stable Diffusion</strong></h3><p>Stable Diffusion has become a leader in AI image generation in recent months. As previously mentioned, the recent breakthrough in Generative AI has resulted from the development of latent diffusion models. <a href="https://huggingface.co/stabilityai/stable-diffusion-2">Stable diffusion</a> is no different. It is an open-source version developed by a company called <a href="https://stability.ai/">Stability AI</a> and its collaborators, CompVis LMU, Runway, EleutherAI, and LAION. The company was founded in 2019 by Emad Mostaque; an ex-hedge fund manager turned tech entrepreneur. In Oct. 2022, Stability AI <a href="https://techcrunch.com/2022/10/17/stability-ai-the-startup-behind-stable-diffusion-raises-101m/">announced its $101M seed round</a>, valuing the company at $1B. Stability AI generates revenue through providing consulting services related to AI as well as its digital service, DreamStudio, which provides access to its image generation tool without the need to install software, obtain technical knowledge, or invest in heavy computing resources.</p><h3>Midjourney</h3><p>Next up, with what also feels like a purposely confusing naming convention, is Midjourney. It’s the company behind a proprietary product using a latent diffusion model, also called Midjourney. Images generated by Midjourney are generally more artistic in style and often look like paintings. The self-funded company was founded by David Holz (who also co-founded <a href="https://en.wikipedia.org/wiki/Leap_Motion">Leap Motion</a>). The difference between Midjourney and other products, such as Stable Diffusion and DALL-E, is that it utilizes the popular social platform Discord as the input tool for creating images. Users join the Midjourney discord channels and interact with the tool using Discord bot commands. This may prove to be a make-or-break strategy for the product compared to Stable Diffusion’s and DALL-E’s browser-based product, as Discord is a contemporary method of interaction that appeals to younger audiences.</p><h3><strong>DALL-E</strong></h3><p>As a seemingly honorific shoutout to the 2008 Disney movie, WALL-E, about a lonely robot in a dystopian future, DALL-E proves to be “painting” a path toward a similar future. Open AI, the company behind DALL-E, was founded by Elon Musk and Sam Altman, with a notable investment of $1B from Microsoft. DALL-E is an early version of image generation, released in January 2021, leveraging a modified version of GPT-3, a generative text machine-learning model. In April 2022, DALL-E 2 was launched using a combination of latent diffusion models and GPT-3. And in November 2022, Open AI released a DALL-E 2 API enabling developers to integrate the tool into their applications. Big tech companies such as <a href="https://techcrunch.com/2022/10/12/microsoft-brings-dall-e-2-to-the-masses-with-designer-and-image-creator/">Microsoft have already begun integrating the API</a> into their products with the launch of their “Canva-killer” app, <a href="https://designer.microsoft.com/">Designer</a>, and an<a href="https://www.youtube.com/watch?v=q74VZizTaEk&amp;t=39s"> Image Creator tool for Bing</a> and Microsoft Edge.</p><h3>Here are some comparisons:</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/994/0*uW244QMK_oF_MuNC" /><figcaption>src: <a href="https://www.marktechpost.com/2022/11/14/how-do-dall%C2%B7e-2-stable-diffusion-and-midjourney-work/">https://www.marktechpost.com/2022/11/14/how-do-dall%C2%B7e-2-stable-diffusion-and-midjourney-work/</a></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/833/0*ykKYh_i3-KvK1jfS" /><figcaption>src: <a href="https://www.marktechpost.com/2022/11/14/how-do-dall%C2%B7e-2-stable-diffusion-and-midjourney-work/">https://www.marktechpost.com/2022/11/14/how-do-dall%C2%B7e-2-stable-diffusion-and-midjourney-work/</a></figcaption></figure><h3><strong>What can we expect to see next?</strong></h3><p>The next 12 to 18 months will be a creative renaissance for a content generation as creators and companies adopt these tools. One powerful aspect of generative AI models is that once they are trained, they can be tweaked and fine-tuned. Take this <a href="https://replicate.com/lambdal/text-to-pokemon">text-to-Pokemon-image model</a>, for example, where you can type in a prompt, and a pokemon-like image is generated. If we ignore the potential for copyright infringement, companies such as Nintendo and Game Freak could benefit from this as well. For instance, The Pokemon Co. has shown signs of creative fatigue with the release of recent games. When Pokemon Sun and Moon was released in 2016, <a href="https://www.theverge.com/2016/6/30/12065976/pokemon-sun-moon-new-characters-bad">The Verge wrote, “these new Pokemon all look terrible</a>.” Using a tool like text-to-Pokemon would enable The Pokemon Co. to rapidly ideate thousands of new Pokemon when designing new game themes and adding more parameters.</p><p>If you’re a brand with lots of IP, you can start to fine-tune these models to help create new concept designs or mash-ups. Companies like Nintendo, Disney, and Netflix will look to leverage this new technology to supercharge their creativity. Furthermore, design products with generative AI features have already started to pop up, such as <a href="https://www.canva.com/features/ai-image-generator/">Canva’s Image Generator</a>, <a href="https://www.figma.com/community/plugin/1145446664512862540/Ando---AI-Copilot-for-Designers">Figma’s Ando — Design Co-Pilot</a>, and <a href="https://designer.microsoft.com/">Microsoft’s Designer</a>, showing an understanding within the design industry that these tools should be embraced rather than scorned. Although most of these tools are still in their early stages, it won’t be long until we begin to see these tools become part of our everyday workflows and industry best practices as the technology matures.</p><p><em>Apply Digital solves complex problems with well-executed digital solutions for some of the most respected brands in the world. We help our clients gain a competitive advantage and delight users with expertly executed digital solutions. Learn more about our services at </em><a href="http://www.applydigital.com/"><em>www.applydigital.com</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7812af8e8a98" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-lab-apply-digital/intro-to-generative-ai-images-stable-diffusion-midjourney-dall-e-7812af8e8a98">Intro to Generative AI — Images (Stable Diffusion, Midjourney, Dall-E)</a> was originally published in <a href="https://medium.com/the-lab-apply-digital">The Lab @ Apply Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>