<?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[Stories by Marc Nehme on Medium]]></title>
        <description><![CDATA[Stories by Marc Nehme on Medium]]></description>
        <link>https://medium.com/@marcnehme?source=rss-5e91026c92ea------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*zNNuhounsqaTNKh_8BLwFA.jpeg</url>
            <title>Stories by Marc Nehme on Medium</title>
            <link>https://medium.com/@marcnehme?source=rss-5e91026c92ea------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 10:50:40 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@marcnehme/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[AI Agents Hype, Explained — What You Really Need to Know to Get Started]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/data-science/ai-agents-hype-explained-what-you-really-need-to-know-to-get-started-e584e38b91be?source=rss-5e91026c92ea------2"><img src="https://cdn-images-1.medium.com/max/1024/1*s2dtl0h7aipYWHfKVC7cUA.png" width="1024"></a></p><p class="medium-feed-snippet">I&#x2019;ll set the record straight&#x200A;&#x2014;&#x200A;AI Agents are not new but advanced. Learn how they&#x2019;ve evolved and where to get started.</p><p class="medium-feed-link"><a href="https://medium.com/data-science/ai-agents-hype-explained-what-you-really-need-to-know-to-get-started-e584e38b91be?source=rss-5e91026c92ea------2">Continue reading on TDS Archive »</a></p></div>]]></description>
            <link>https://medium.com/data-science/ai-agents-hype-explained-what-you-really-need-to-know-to-get-started-e584e38b91be?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/e584e38b91be</guid>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[large-language-models]]></category>
            <category><![CDATA[llm]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Mon, 06 Jan 2025 21:51:08 GMT</pubDate>
            <atom:updated>2025-01-06T21:51:52.739Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Revolutionizing Call Center Training With AI: How Advanced Voice Transforms Agent Development]]></title>
            <link>https://marcnehme.medium.com/revolutionizing-call-center-training-with-ai-how-advanced-voice-transforms-agent-development-c93c5829c8fd?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/c93c5829c8fd</guid>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[speech]]></category>
            <category><![CDATA[call-center]]></category>
            <category><![CDATA[openai]]></category>
            <category><![CDATA[chatgpt]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Mon, 04 Nov 2024 14:24:43 GMT</pubDate>
            <atom:updated>2024-11-11T15:35:43.188Z</atom:updated>
            <content:encoded><![CDATA[<h3>GPT-4o Revolutionizing Call Center Training: How Advanced Voice Transforms Agent Development</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ah8B3zLrGVwKNytkJmn12A.jpeg" /></figure><p>Call center has always been one of the top AI use cases for enterprise businesses. When we talk call center scenarios, we typically think of “agent assist” <em>(AI helping human agents on customer calls)</em> or “customer assist” <em>(AI helping customers through a conversational bot), </em>however what if there was a way to get ahead of that and better train human agents before they ever take their first customer call? GPT-4o has now created the opportunity for another innovative way to help increase productivity, improve handle time and customer satisfaction in the call center.</p><p>In the fast-paced world of customer service, the quality of call center training can make or break a company’s reputation. Traditional training methods like classroom sessions and shadowing experienced agents, while valuable, often leave new agents feeling unprepared for real-world customer interactions. Over the past decade, I’ve worked with numerous enterprise customers in the call center space. Having also worked in a call center during my college years, I’ve witnessed this firsthand.</p><p>Enter AI and <strong>advanced voice mode</strong> — the game-changing innovation to prepare call center agents for success. You can access this capability using the Realtime API with <a href="https://learn.microsoft.com/en-us/azure/ai-services/openai/realtime-audio-quickstart?pivots=programming-language-ai-studio">Microsoft Azure OpenAI</a> or <a href="https://openai.com/index/introducing-the-realtime-api/">OpenAI</a>.<em> (Note that as of this post date, both are still in beta/preview). </em>There are many different applications of how this technology can enhance businesses, side hustles, and every day life. See my other post on <a href="https://medium.com/@marcnehme/meet-your-new-co-host-how-ai-will-transform-podcasting-bc6c21e0844c">How AI Will Transform Podcasting</a> which also discusses key capabilities, results, optimization, and things to be mindful of when using this advanced voice technology. In this post, I will focus on how this will advance call center training.</p><h3>The Training Gap in Traditional Call Centers</h3><p>Most call centers follow a familiar training pattern: agents study protocols in classrooms, take learning modules, memorize scripts, and shadow experienced staff. While these methods build a foundation, they share a common limitation — the lack of hands-on practice in a risk-free environment. New agents often face their first real customer interactions with anxiety, potentially leading to suboptimal service and decreased customer satisfaction. You do not want to sacrifice customer service at the expense of training that can be mitigated.</p><h3>Advanced Voice Mode: The Game Changer in Agent Training</h3><p>Imagine a training environment where new agents can practice handling countless customer scenarios without the pressure of real-world consequences. This is exactly what advanced voice delivers. It can simulate realistic customer interactions with specific scenarios and different personalities to allow agents to get “live” experience. Here are some <strong>key benefits</strong>:</p><ol><li><strong>Risk-Free Learning Environment </strong>— agents can make mistakes without impacting real customers, reducing stress, allowing them to have better learning outcomes, and building their confidence before taking real customer calls</li><li><strong>Diverse Scenario Coverage</strong> — exposure to a wide range of customer scenarios, personalities, emotions, communication styles, and multiple languages</li><li><strong>Consistency &amp; Availability</strong> —standardized scenarios to practice common customer issues or asks, ensuring consistent learning experiences, with training available 24/7 which is key for multi-timezones</li></ol><p>Below, you can listen to my demonstration of a trainee-AI customer interaction <em>(I am acting as the call center agent trainee)</em>. For this demo, I used the <a href="https://learn.microsoft.com/en-us/azure/ai-services/openai/realtime-audio-quickstart?pivots=programming-language-ai-studio">Azure OpenAI gpt-4o </a>model in <a href="https://learn.microsoft.com/en-us/azure/ai-studio/what-is-ai-studio">AI Studio</a>. <em>(Fans of the Netflix series “Breaking Bad” will especially enjoy this demo!)</em></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FJ-Lg1ckhi60%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DJ-Lg1ckhi60&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FJ-Lg1ckhi60%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/cbeb7a5804e92364446c9fcc39e13634/href">https://medium.com/media/cbeb7a5804e92364446c9fcc39e13634/href</a></iframe><h3>Where Do You Start?</h3><p>Below are some tips in getting started. This is not a complete list of everything you need to do, but are some of the major items that are critical to implement.</p><ul><li>Evaluate current training pain points</li><li>Identify specific scenarios for your business with associated conversation flows</li><li>Identify the different personalities/profiles of your to-be AI customers</li><li>Create the prompts/context for each AI customer profiles to handle each of the scenarios</li><li>Have expert call center agents do extensive testing of the AI-based customers prior to exposing this to the trainees</li><li>Set clear objectives and KPIs to identify what success means</li><li>Use this AI capability responsibly <em>(see </em><a href="https://www.microsoft.com/en-us/ai/responsible-ai"><em>Microsoft Responsible AI principles</em></a><em>)</em> and ensure content safety to ground the AI-based customers to stay focused on the business scenario</li></ul><h3>The Future of Call Center Excellence</h3><p>This advanced voice mode technology, provided by Microsoft Azure OpenAI and OpenAI, represents a paradigm shift in call center training. By providing a safe, focused practice environment, it bridges the gap between theoretical knowledge and practical application. Don’t let your agents face their first customer interaction unprepared. Organizations that embrace this technology today will be better positioned to deliver exceptional customer service tomorrow.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c93c5829c8fd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Meet Your New Co-Host: How AI Will Transform Podcasting]]></title>
            <link>https://marcnehme.medium.com/meet-your-new-co-host-how-ai-will-transform-podcasting-bc6c21e0844c?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/bc6c21e0844c</guid>
            <category><![CDATA[openai]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[podcast]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[speech]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Mon, 28 Oct 2024 13:58:02 GMT</pubDate>
            <atom:updated>2024-10-28T15:59:11.681Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SVO7DI4A37zYlU1MVrGVjA.jpeg" /><figcaption>(image generated by OpenAI gpt-4o)</figcaption></figure><p>Imagine a feature so versatile it can empower businesses, elevate side hustles, and spark creativity for hobbyists — all while making everyday tasks smoother and smarter. I’ve been diving into this game-changing capability, and the potential is nothing short of remarkable. In this post, I’ll reveal one of its most groundbreaking applications and show how it’s set to redefine what’s possible in podcasting.</p><h3>Context</h3><p>OpenAI just recently released the “advanced voice mode” capability in ChatGPT as well as the realtime API, which allows for speech-to-speech conversations. This was also released on Microsoft Azure through AI Studio with the <a href="https://azure.microsoft.com/en-us/blog/announcing-new-products-and-features-for-azure-openai-service-including-gpt-4o-realtime-preview-with-audio-and-speech-capabilities/">GPT-4o-Realtime-Preview</a> model.</p><p>Let’s dive into why this new capability is such a breakthrough in AI technology. Previously, to achieve speech-to-speech like capability, one would need to stitch together multiple steps. First, you would need to transcribe the spoken audio to text using a speech-to-text service to. Then run that text through an analysis model to generate a response. Finally, you’d convert the model’s response back to speech using a text-to-speech service. This multi-step process is more tedious to implement and can lead to decreased quality, such as longer latency and a loss of emotion.</p><p>With the new realtime API, everything flows smoothly and intuitively. It allows you to stream audio inputs and outputs directly, enabling more natural conversational experiences, and all in ONE API call. You can read more about the announcement and features <a href="https://openai.com/index/introducing-the-realtime-api/">here</a>.</p><h3>How AI Can Be the Ultimate Co-Host for Your Podcast</h3><p>Now, let me get down to brass tacks. I’m not only an AI enthusiast and practitioner, but I’m also a hardcore professional wrestling fan. As a kid, <em>(and maybe some as an adult) </em>I’ve role-played wrestlers cutting promos <em>(monologues to promote themselves)</em> or commentators calling wrestling matches. I also listen to several wrestling podcasts, such as <a href="https://www.youtube.com/@SomethingToWrestle">Something to Wrestle with Bruce Prichard</a> and Conrad Thompson <em>(a must listen for any wrestling fan!)</em> As I was listening to their podcast, I was thinking of how cool it would be to have AI be the co-host of a podcast, leveraging the new advanced voice capability.</p><p>You can evaluate this on Microsoft Azure with the Real-Time audio playground in <a href="https://azure.microsoft.com/en-us/products/ai-studio">Azure AI Studio</a>, however for this test, I used OpenAI’s ChatGPT app on my iPhone with gpt-4o. I was really impressed with the results that I have shared below, as well as my demo. <em>Note: at the end of the demo, I was amazed with the questions “Becky” created on her own as I did not guide it on that, and took those on the fly!</em></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F1ha-uxpjMno%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D1ha-uxpjMno&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F1ha-uxpjMno%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/2595de404dca46819568f28aab4de50c/href">https://medium.com/media/2595de404dca46819568f28aab4de50c/href</a></iframe><h3>Results &amp; Optimization</h3><p>Here are some insights and best practices that I learned to drive positive results as an AI co-host for a podcast…</p><p><strong>The conversational nature of ChatGPT was phenomenal</strong>. In my prompt prior to starting the podcast, I provided context to ChatGPT to ensure it provided more optimal results <em>(this was not included in the demo above)</em>. Below are some things I implemented:</p><ol><li><strong>Give it a name</strong> — I gave my ChatGPT a name to sound more realistic. I called it “Becky”, paying homage to one of the most popular and successful women’s wrestlers of all time — Becky Lynch.</li><li><strong>Set the stage — </strong>I provided very specific context to Becky of what the goal was (simulating a live podcast) and what I needed Becky to do. For example, I specified the topic of wrestling in the late 1980s, and that we’d be referencing Hulk Hogan and Macho Man’s rise to prominence and downfall as a team.</li><li><strong>Provide specifics on speaking — </strong>I told Becky to sound realistic and not provide perfectly spoken responses, to be casual, to show emotion, and not provide long answers. <em>(You can also select from one of the preset voices that you would like to use. I selected “Sol”. There are currently 9 different options.)</em></li><li><strong>Have your Co-host ask you questions — </strong>I told Becky to ask me questions as a way to have a natural conversation on a specific topic. This was to help ensure that the podcast does not turn into a one way conversation of only me asking Becky questions.</li></ol><p><strong>Model Knowledge was great</strong> — this was very impressive. The level of detail that Becky was able to provide with hints or inferences were spot on. Becky used terms like “heel” (which means “bad guy” in wrestling speak). I did not need to be very explicit in guiding Becky during the actual podcast and that made the result sound alot more natural.</p><p><strong>Creativity</strong> — when asking Becky about where the Wrestlemania 5 main event match ranks in Macho Man’s career all time, Becky weaved in a historical reference to his Wrestlemania 3 match, which is also regarded as one of the greatest matches of all time. This proved Becky’s comments to be very intelligent and relevant.</p><p><strong>Handling interruptions </strong>— this capability of allowing me (user) to interrupt Becky (ChatGPT) adds alot to the reality of how podcasting plays out. I was able to interrupt Becky mid sentence, and Becky would adapt to whatever direction I’ve gone in.</p><p><strong>Multi-person Podcast </strong>— In my demo, it was just myself and Becky, however you can do this with multiple humans along with Becky. This would be even more engaging where Becky would listen in and naturally contribute to the conversation.</p><h3>Things to be mindful of:</h3><p><strong>Pausing — </strong>There was a slight pause <em>(~1–2 sec) </em>between when I finished a sentence and when Becky spoke. Its very reasonable as I did this from my iPhone. I also did some testing through Azure AI Studio and the response times were similar but can be made faster with the silence duration parameter. Subsequent tests I ran were also faster. There are many factors that I will not get into here that can affect latency but wanted to share my results at the time of my demonstration. This will vary.</p><p><strong>Longer answers — </strong>Some answers were still a bit long despite me specifying this as context in the prompt. I’ll need to play with this a little more to get the right depth of response for my use case.</p><p><strong>Repetitive phrases</strong> — Becky loved to say “What do you think, Marc?” quite a bit. This can be tweaked with parameter tuning and additional prompt techniques, so totally solvable.</p><p><strong>Optimize for your Use case </strong>— in my example and the first time I tested this, I found that Becky would give away the outcome of the match in the first piece of feedback provided. I added some instructions in the prompt to tell Becky not to give away the outcome of the matches right away. I instructed Becky to have atleast one back and forth chat before discussing the outcome, unless I specified otherwise. It worked well.</p><h3>Conclusion</h3><p>We’ve only begun to explore the potential of the advanced voice mode capability. Having AI as a podcast co-host, leveraging advanced voice mode or the realtime API, opens many new possibilities. It is cutting edge technology. Imagine a co-host that could appeal to all audiences, discuss nearly any topic, with a plethora of knowledge, and interact in a natural conversational style — in real time. It will transform the podcasting game with infinite captivating, interactive and entertaining experiences for years to come.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bc6c21e0844c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Key ingredients to a successful partnership]]></title>
            <link>https://marcnehme.medium.com/key-ingredients-to-a-successful-partnership-8bca6730f5a0?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/8bca6730f5a0</guid>
            <category><![CDATA[strategy]]></category>
            <category><![CDATA[alliances]]></category>
            <category><![CDATA[partnerships]]></category>
            <category><![CDATA[collaboration]]></category>
            <category><![CDATA[best-practices]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Thu, 21 Jan 2021 19:06:38 GMT</pubDate>
            <atom:updated>2021-01-21T19:09:19.810Z</atom:updated>
            <content:encoded><![CDATA[<h4>How to win together</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SFP4BsPBV1LAcyAByUfplg.jpeg" /></figure><p>In a previous <a href="https://marcnehme.medium.com/why-partnerships-have-become-critical-to-all-companies-c59f5bea0823">blog</a>, I explained why partnerships are critical to the success of organizations who want to continuously scale. However, there are some recipes to follow to achieve success. With over a decade of experience in working with strategic partners and customers, I will share my perspective on the key elements that lead to successful partnerships.</p><h3>“One size fits all” does not work</h3><p>It’s very important to understand that trying to use one model for partnerships does not work very well. You should have different partnership models that you offer. Some popular examples are OEM, co-development and co-sell. There are also hands-on and hands-off models. There are many different models that exist and can be created. You have to understand each partner’s position to properly classify them into a partnership model in your organization. If you do not have an existing model that fits a specific partner, do not force them into a model. You should consider creating a new one to accommodate.</p><h3>Transparency</h3><p>This is one of the most critical, hidden gems in any partnership. While this seems like a given, it is not as obvious as one may think. To be successful in a partnership, you need to be <strong>transparent</strong> with each other. This is something that I have experienced great success with. I’ve always been very transparent in partnerships. Address all the elephants in the room and peel back the onion.</p><p>From the start, <strong>be open with your goals and objectives</strong>. It’s easier for both sides to march towards success if you understand where the other side wants to land. It helps craft the right messaging and direction going forward. If you have <strong>competing technology</strong>, that it fine. In fact, it’s common in the tech industry. Highlight it from the beginning. Focus on areas of how you can <em>compliment</em> each other versus areas of <em>competition</em>. Talk about it.</p><p>In your initial meetings, discuss your organization’s mode of operation. Explain your company culture. I’ve learned a great deal about the different cultures of many partners. It’s very interesting and educational. You may even find yourself adopting some of their practices to improve your organization! Be transparent with your priority of the partnership to set the right expectation. This will help avoid alot of misunderstanding and over communicating down the road. It will ensure better alignment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Na5TbCa-tT0U1h_8xWk5wA.jpeg" /><figcaption>Photo by Kraken on Unsplash</figcaption></figure><h3>Investment</h3><p>There needs to be “skin in the game” on both ends. Lop-sided partnerships do not turn out all that well, for either side. Partnerships that have the greatest success are the ones that have significant and consistent engagement on both sides. Clear expectations should be set and agreed upon. Roles need to be clearly defined. Make sure you have the resources you need to support your partnership model. If you do not have that in place, raise it to your leadership team right away. The earlier you identify a need, the sooner that it can be addressed.</p><p>It’s very encouraging for both sides to contribute some progress, early on and often. That could be data, code, resources, securing pilot clients, etc. If you’ve had many meetings with several weeks gone by and are still trying to align on the basics, then this may be a signal to investigate your approach. Gone are the days of 9+ month deployments. Being 2021, customers expect to see value in weeks to a few months. And I completely support that. The point is, you need to get the partnership activities moving fast. This is so you can capitalize on the moment, stay ahead of the market and product successful results.</p><h3>Alignment</h3><p>It is imperative that you work together with your partner as <strong>one alliance team</strong>. The topics above will help get you there. Align on one strategy, one plan, and realistic expectations. Make sure both sides understand those respective expectations. Build strong relationships with each other. Both sides should be singing the same song. This makes for a healthy and more successful partnership.</p><p><strong>Marc</strong> is<strong> </strong>the CTO for the IBM Watson AI Strategic Partnerships organization. When not leading the technology strategy and vision for IBM Partnerships, Marc enjoys DJing, playing video games and wrestling. Learn more @ <a href="https://marcnehme.com/">MarcNehme.com</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8bca6730f5a0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Where do I start with AI?]]></title>
            <link>https://medium.com/data-science/where-do-i-start-with-ai-c2769c61fcde?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/c2769c61fcde</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[getting-started]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[best-practices]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Fri, 08 Jan 2021 01:18:49 GMT</pubDate>
            <atom:updated>2021-01-11T17:48:06.877Z</atom:updated>
            <content:encoded><![CDATA[<h4>Entering the world of artificial intelligence</h4><p>Over the last 7+ years, I&#39;ve worked with many leaders who were looking to embark on their first AI journey. Most of them did not know how or where to start. Some thought they did. In order for your business to realize the gains possible using AI, you need to make sure you start your journey on the right path. Here are some key things to consider when beginning your journey…</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NbDQDc_aLx5E_7dwAHME0g.jpeg" /><figcaption>Photo by Vlad Bagacian on <a href="https://unsplash.com/">Unsplash</a></figcaption></figure><h3>Goal</h3><p>You need to be able to clearly articulate what your ultimate goal is. If you cannot do that, then you are not ready to actually start the journey. Your goal should NOT be “I want to run an AI project”. That is too vague and not outcome or results driven. That is a red flag and often a recipe for failure.</p><p>Clear goals can also be at different levels. A targeted, yet somewhat high level goal could be “I want to improve the customer experience”. The general outcome is clear — “improve the customer experience” but it is still vague in the sense of there are no dimensions to the goal. A more specific goal would be “I want to improve the customer experience in the call center”. A very specific goal would be “I want to improve the customer experience in the call center by reducing call handle times”. The more specific you can get, the better off you will be downstream.</p><p>Again, attempting to start a project using AI without a clear vision and goal will likely not provide you with optimal results. Do not try to implement AI just to “do it”. Start the journey with specific goals in mind.</p><h3>Identifying areas of optimization</h3><p>The next step is identifying where the pain points are. In this example, you’ll look at the end to end customer experience. You’ll identify what is optimal and non-optimal, the as-is and to-be scenarios, the actors involved, and more. This can be achieved through design thinking sessions — I will not get into that topic here. There are a lot of online resources that discuss and explain design thinking. IBM has a phenomenal design thinking program to help customers get started on their journey.</p><p>Let’s assume you go through the design thinking session and have those areas identified. In that session, you’ll also identify the AI capabilities to optimize those areas. For example, implementing a conversational AI assistant will help your call center agents get answers to questions faster, which will reduce call handle times. Always identify the use case/area of optimization first THEN map technology to solve it. Do not do the opposite!</p><h3>A big decision</h3><p>You also need to decide how you will handle the implementation. There are several choices here.</p><h3>1 — Build your own AI practice and develop your own AI technology</h3><p>This will often require the most time and investment to get going. As a leader, you need to assess many things such as: Do you have existing AI skills internally? If not, how will you acquire them? What technology do you have available to you to get started? How long will this all take compared to your timeline? Will this deliver the best results? Does this align with our company strategy and vision? Being 2021, there are many companies out there who already have mature AI platforms. You need to assess if reinventing the wheel for your own purposes is necessary.</p><h3>2 — Build your own AI practice using an existing AI vendor platform</h3><p>Here, you up-skill your employees on another AI vendor’s platform, such as IBM Watson or Microsoft AI. In this option, your bigger investment is in your own employees knowledge. The benefit here is that you do not need to develop the AI technology — just learn how to use it. Your employees would customize the virtual assistant, in this example, to satisfy your use case. For example, using IBM Watson Assistant (an AI-powered virtual assistant), you can leverage the existing cloud service APIs (via SaaS model). You can also deploy Watson in your own cloud on Azure, AWS or GCP, or you can deploy Watson on premise if you so choose. In either case, your team can focus on building the virtual assistant to meet your use case. They do not need to worry about developing or maintaining the AI service itself.</p><h3>3 — Hire professional services to do the implementation for you</h3><p>As a leader, there are situations where you simply do not have the time or staff to go with options 1 or 2. AI can also be daunting depending on your skill set. You can choose this option of hiring professional services to implement the virtual assistant for you. Here, you still leverage an existing AI platform in the market, but you will hire a company to tailor the assistant to fit your use case. This should give you some peace of mind since you have hired professionals. They will guide the implementation every step of the way. Your main involvement will be to provide the domain subject matter expertise to the implementation team.</p><h3>4 — Build your own AI practice and hire professional services</h3><p>This is a hybrid model of options 2 and 3 above. You can hire professional services to guide the implementation while your internal team up-skills side-by-side with the professionals. There are many benefits to this approach. The main benefit is you get the professional assistance you need along the way while training your own staff to be self-sufficient. Here, you balance the main project deliverable of a virtual assistant along with up-skilling your team. You&#39;re accomplishing two major goals, simultaneously.</p><p>There are many aspects to be considered when making your choice. You need to assess budget, timeline, strategy, staff, etc. There are also different ways you can take your journey. For example, you can start with option 3 for your first release and then pivot to option 4 for the second release. Ideally, you want to pick one strategy and try to stick to it to maximize value.</p><h3>In the end</h3><p>The bottom line is you need to ensure you are starting your AI journey for the right reasons, with quantifiable goals, meaningful business outcomes and the right implementation plan. Seek professional help if you are unsure. There are also many online resources to help educate you and get you some hands-on with AI, for free, such as <a href="https://cloud.ibm.com/catalog?category=ai#services">IBM Watson services</a>.</p><p><strong>Marc</strong> is<strong> </strong>the CTO for the IBM Watson AI Strategic Partnerships organization. When not leading the technology strategy and vision for IBM Partnerships, Marc enjoys DJing, playing video games and wrestling.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c2769c61fcde" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/where-do-i-start-with-ai-c2769c61fcde">Where do I start with AI?</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Hidden gems in AI deployments]]></title>
            <link>https://medium.com/data-science/hidden-gems-in-ai-deployments-53d8d38b3403?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/53d8d38b3403</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[best-practices]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[office-hours]]></category>
            <category><![CDATA[editors-pick]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Wed, 09 Dec 2020 12:02:24 GMT</pubDate>
            <atom:updated>2020-12-15T22:07:29.870Z</atom:updated>
            <content:encoded><![CDATA[<h4><a href="https://towardsdatascience.com/tagged/office-hours">Office Hours</a></h4><h4>Things that are critical to your success in deploying AI</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2b_jyP7R65_s1ccb89nX2Q.jpeg" /><figcaption>Photo by Michael Dziedzic on <a href="https://unsplash.com/">Unsplash</a></figcaption></figure><p>Since 2014, I’ve been leading and advising teams to design and deliver AI-powered solutions from concept to production. While the technology has evolved significantly, the recipe for delivery has largely remained the same. I want to share some best practices and critical aspects that are often underestimated or overlooked. These are the “hidden gems” that I have found to be critical to the success of a production deployment.</p><h3><strong>ROI &amp; </strong>Success Criteria</h3><p>The first thing both you and your customer need to understand and have complete clarity on is the purpose of the project. Yes, you need to define the use case, scenarios, personas, and more. But you need to answer several questions, such as: What are the actual outcomes the customer is looking to achieve? How will this help improve the customer’s business?</p><p>I have seen many AI deployments where the outcome was a technical success but a business failure. Hence, those projects were not successful. The technical aspects of the solution were working correctly but there was no quantifiable business value. The ROI could not be proven, for example, no additional revenue generated or little to no cost savings.</p><p>That is why it is critical to understand and define success criteria for any project. This should be created and agreed to, by both parties, before you start. This helps ensure that the outcome is quantifiable and you can definitively state if the project was a success, or not. Here are some examples of success criteria statements:</p><ul><li><em>“Meet an accuracy level of 70% or above for all functional tests, using an agreed upon test rating system”</em></li><li><em>“Increase an employee’s efficiency rate by 30%”</em></li><li><em>“Improve the system processing time by at least 20%”</em></li></ul><h3>Scope</h3><p>This is another item that needs to be clearly defined and documented. Every project needs a scope — that is, what is supposed to be implemented during the actual delivery of the project. This captures what is “in bounds” and also should outline some related or common “out of scope” items. Here are some examples:</p><ul><li><em>“Will integrate with one PostgreSQL db to extract all financial data”</em></li><li><em>“Integration with any other data source, other than PostgreSQL, is out of scope”</em></li><li><em>“Will not exceed modeling more than 200 unique answers”</em></li></ul><p>Documenting your scope will help you avoid chasing a moving target. This allows you to state that you have implemented everything promised in the original scope of work. This also will help identify and avoid scope creep — where new items are added to the project during delivery.</p><h3><strong>Understanding the Requirements</strong></h3><p>This is another very critical part of the process. You must listen to your customer, listen to their challenges, listen to where they want to be. Not all customers are knowledgeable in the AI space. Their experience can range from having some hands-on experience to knowing nothing at all. That is why you are in the picture — to help guide them along this journey!</p><p>Document your requirements. Mutually agree to those requirements. Show how those requirements map to the solution’s use case, scope and success criteria. You can partition the requirements into two main types, functional and non-functional. Below are some examples.</p><p>Functional requirements:</p><ul><li><em>“Must support all products mentioned in the Form 101 document”</em></li><li><em>“Must provide a confidence level with each answer”</em></li><li><em>“Must provide a feature to allow the end user to speak to a live agent”</em></li></ul><p>Non-functional requirements:</p><ul><li><em>“Must support 1000 concurrent users”</em></li><li><em>“Must support mobile browsers (Chrome, Safari) and desktop browsers (Firefox and Chrome)”</em></li><li><em>“Must provide a response to users within 5 seconds or less”</em></li></ul><p>Without clearly defined requirements, you risk misalignment with your customer. You risk having a lack of clarity, both internally and externally. You need to understand the requirements to generate epics, stories and tasks to assign the work to your teams and to track progress.</p><p>ROI, success criteria, scope and requirements can be discussed and vetted via design thinking sessions <em>(or other focused sessions).</em> But they do not end there. They need to be formalized into mutually agreed upon documentation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jRq4qe1D3Ujh8kLhHgMfRQ.jpeg" /><figcaption>Photo by Gabrielle Henderson on <a href="https://unsplash.com/">Unsplash</a></figcaption></figure><h3>Data</h3><p>In order to properly solution any AI-powered deployment, you need to see and understand the data that you will be working with. Always ask for samples of the actual data that will be used as early as you can. Assess the data. There are many questions you need to answer, such as:</p><ul><li><em>Does your customer have access/privileges to all the data you need?</em></li><li><em>Are there any security hurdles or compliances, on either end, that need to be worked through?</em></li><li><em>Is the data ready “as-is” to be used for your project? If it is not, then what needs to be done to get the data into a consumable format?</em></li><li><em>Do you have “enough” data? That is a very subjective concept and there is no one size fits all. You need to assess to make sure you have a proper distribution of the data?</em></li></ul><p>I’ve seen many projects that have been delayed, failed or never took off due to data issues. This is definitively a showstopper so make sure you have assessed the data!</p><h3>The Solution</h3><p>A common mistake when starting a project, is automatically assuming the technology to be used. Do not do this. Do not go into a project with any firm assumptions of which technology you want to use. First, you need to completely understand what the customer is looking to achieve <em>(all the things I have discuss above). </em>Then, you can identify the appropriate technology in your architecture to meet the project’s needs.</p><p>The technology you propose should be used for its intended purpose. Most products are designed with specific use cases and certain data in mind. This can also affect what type of product support you receive. Make sure to understand the product limitations before incorporating it into the overall architecture. A deciding factor is, does the technology produce results that directly map to what my customer has asked for?</p><h3>Playbacks</h3><p>Put something in front of your customer early and often. Even if its just wireframes or very early development work. This will help save lots of time and can be the difference between a failed or successful project. The sooner you put something in front of your customer, the better. This will confirm your understanding of what you need to be building. It confirms if you‘re on the right, or wrong, path. Playbacks can be at held at whatever interval you choose. Two weeks is a typical interval most of my teams have used successfully.</p><h3>Testing</h3><p>You want to test your solution as early as possible. You also want to ensure that the people testing your solution are the real end users (or the closest thing to it). For example, let’s say your end user persona is targeted to be new employees only. Do not have SMEs or seasoned employees be the primary testers of your solution. It is not representative and will not be providing you contextual feedback. While it is important for those experts to test and provide their feedback, they will not be the ones using the solution. At the end of the day, it is the new employee persona that will be using this solution. Their opinion should carry more weight.</p><p>The number of test cycles you go through will depend on various factors such as project timeline, quality of results, scope, etc. In my teams’ experience, we have found that on average, three full test cycles (no less) are both effective and realistic with all things considered. The more you test, the more you can enhance the solution. You will eventually hit a point of diminishing return.</p><h3>The Bottom Line</h3><p>There are many different aspects to making an AI project a reality and success. I did not discuss every single aspect above but I have shared my experience and lessons learned for some of the most critical activities. These are the areas that I have seen to be the most common places where teams have struggled and succeeded. Ultimately, how you approach and complete these activities will dictate the level of success of your projects.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hq4Uk6BFgc_0E_DVVcsavg.jpeg" /><figcaption>Photo credit — <a href="https://unsplash.com/">Unsplash</a></figcaption></figure><p><strong>Marc</strong> is<strong> </strong>the CTO for the IBM Watson AI Strategic Partnerships organization. When not leading the technology strategy and vision for IBM Partnerships, Marc enjoys DJing, playing video games and wrestling.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=53d8d38b3403" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/hidden-gems-in-ai-deployments-53d8d38b3403">Hidden gems in AI deployments</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</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 Manage Multiple Languages with Watson Assistant]]></title>
            <link>https://medium.com/ibm-data-ai/how-to-manage-multiple-languages-with-watson-assistant-a8eb0436d348?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/a8eb0436d348</guid>
            <category><![CDATA[tutorial]]></category>
            <category><![CDATA[nlp]]></category>
            <category><![CDATA[ibm-watson-ai]]></category>
            <category><![CDATA[multilingual-chat]]></category>
            <category><![CDATA[conversational-ui]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Mon, 03 Aug 2020 11:55:52 GMT</pubDate>
            <atom:updated>2021-05-26T00:41:57.898Z</atom:updated>
            <content:encoded><![CDATA[<h3>Multilingual Management with Watson Assistant</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9tQW1OcsjKmyk7VLaOleFg.jpeg" /></figure><p>It is very common to implement a conversational assistant in multiple languages. Let’s say you have built a customer service assistant. One of the requirements may be to support additional languages, such as Arabic and Spanish, to accommodate your customer base. Often, I am asked how best to manage multiple skills, formerly called workspaces, in <a href="https://www.ibm.com/cloud/watson-assistant/">IBM Watson Assistant</a> to achieve this goal. Though complete automation is a non-trivial task, there are ways this can be managed with some automation. Below are some suggestions to get you on your path.</p><p>Let’s assume you have built your Watson Assistant in the English language. Now you need to build the Arabic version of the assistant. The first thing you should do is review the list of <a href="https://cloud.ibm.com/docs/assistant?topic=assistant-language-support">supported languages</a> and features for your languages of interest. Each language does not support all features. For example, contextual entities are supported in English but not in Arabic. So just be mindful of those feature differences and account for them accordingly.</p><p>Next, you need to export your English skill from Watson Assistant. Before you do that, you want to make sure that the “Irrelevance detection” option is set to “Keep existing”. There is an enhanced capability to detect irrelevant topics, however as of this post, it is only available in English. If this is not set properly, the import of your skill will throw an error later.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/641/1*WQy06xJgfozhi2ZUr7oGWA.png" /><figcaption>Irrelevance detection option</figcaption></figure><p>Now you are ready to export your English skill. Go to your “Skills” section in Watson Assistant, click on the ellipsis and then select “Export”. This will export your skill to a json file.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/436/1*sQfpbtclPWpk9OJrWo6kug.png" /><figcaption>Export your dialog skill</figcaption></figure><p>Open that json file in your text editor of choice and look for the language parameter. Change the ‘en’ to the target language of your choice. In this case, ‘ar’ for the Arabic language. You can find the appropriate value for your language <a href="https://cloud.ibm.com/docs/assistant?topic=assistant-language-support">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/728/1*pxRZ1dd60GTlphEOwSg62w.png" /></figure><p>Once you have made the change, save the json file. Go back to Watson Assistant, create a new dialog skill and select the “Import skill” option. Then point to your newly modified json file, which contains the Arabic language value. Watson Assistant will build your new skill with the core language set to Arabic.</p><p>To be clear, what you now have, is a copy of your English skill. It has all the dialog, intents, etc, still in the English language. But it is inside the context of an Arabic skill according to Watson Assistant. Note that there are additional <a href="https://cloud.ibm.com/docs/assistant?topic=assistant-language-support#language-support-configure-bidirectional">bidirectional language preferences</a> that you can optionally configure in an Arabic skill.</p><p><strong>Translate Content</strong></p><p>The next step is to translate all your textual content from English to Arabic. There are two major steps in this process. The first step is getting the content translated. The second step is updating the skill with the translated content. There are several ways you can go about doing this.</p><p>One automation option is to use <a href="https://www.ibm.com/cloud/watson-language-translator">IBM Watson Language Translator</a> to translate all your text from English to Arabic. You would need to write a script to extract the relevant text from your English skill’s json file to process those through the Watson Language Translator service. <em>Those results need to be stored somewhere such as a data store or a temporary location — I’ll discuss this in the next section. </em>Remember, this is a conversational use case. English and Arabic are wildly different languages in their form and origin<em> (I speak Arabic, so I know firsthand :) .</em> I would highly suggest having a native Arabic resource <em>(who also speaks, reads and writes English fluently) </em>review all the translations to ensure that the context is optimal for conversation.</p><p>Another option would be to have a native Arabic resource manually translate the content. You would need to create a data store, such as a spreadsheet, that contains all your dialog structure, intents, etc. The resource can use this to review the content and add their translations. Using an external data store will make it much easier to manage your skills for purposes like this.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IoqhqksPyD4jW8UezzuXdQ.jpeg" /></figure><p><strong>Updating Content</strong></p><p>Now that you have the content translated, the ideal scenario would be to trigger a process to have all the content automatically updated in your skill, in Watson Assistant. This is certainly achievable but requires some work to create a program that will take the new Arabic translations from your data store or temporary storage and update the English equivalents. You can programmatically apply the updates to your skill json file. You need to have an organized and detailed data store that contains a key reference <em>(eg dialog node ID)</em> to ensure this is done smoothly.</p><p>Another option is to do this manually in Watson Assistant. It would not be the most exciting task, yet it will get the job done. The resource who updates the content in Watson Assistant will still require the data store reference to know how to appropriately update the content.</p><p>The option that you decide on will depend on several factors such as your sizing of effort, resources available, etc.</p><p><strong>Managing Content Going Forward</strong></p><p>Once you define your process to translate your skills, you can use that approach to manage any updates going forward. There are other possible options out there to accomplish these tasks but these tips will help get you started on your path.</p><p><strong>Marc</strong> is<strong> </strong>the CTO for IBM Watson AI Strategic Partnerships. When not leading the technology strategy and vision for IBM Partnerships, Marc enjoys DJing, playing video games and wrestling.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a8eb0436d348" width="1" height="1" alt=""><hr><p><a href="https://medium.com/ibm-data-ai/how-to-manage-multiple-languages-with-watson-assistant-a8eb0436d348">How to Manage Multiple Languages with Watson Assistant</a> was originally published in <a href="https://medium.com/ibm-data-ai">IBM Data Science in Practice</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Document Governance with Watson Discovery and Watson Assistant]]></title>
            <link>https://medium.com/ibm-data-ai/document-governance-with-watson-discovery-and-watson-assistant-9b6e1edd1e08?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/9b6e1edd1e08</guid>
            <category><![CDATA[tutorial]]></category>
            <category><![CDATA[ai-trust]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[nlp]]></category>
            <category><![CDATA[conversational-ui]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Thu, 27 Feb 2020 18:21:44 GMT</pubDate>
            <atom:updated>2021-05-21T22:22:28.764Z</atom:updated>
            <content:encoded><![CDATA[<h3>Document Governance with Watson Discovery for Watson Assistant</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JLhE42bY1mQ6z_MEYcuBKQ.jpeg" /></figure><p>You have many teams or roles within your organization. They all are using the same virtual assistant to get answers to their questions. Your virtual assistant handles common dialog and FAQs for all your teams. Yet, you need to ensure that each team is getting the correct answers, from the right documents. These document-based answers could be based on <em>“long tail” </em>or less frequently asked questions.<em> </em>Or the answers may require documentation to follow, such as steps of a process or a list of guidelines. You do not need to imagine too hard as this is a very common scenario.</p><p><em>In this read, I share two approaches to how you can achieve document governance with Watson Discovery along with Watson Assistant.</em></p><p>Please note that my guidance here is for those who are using IBM Cloud public. If you are using “IBM Cloud Pak for Data”, then “<a href="https://cloud.ibm.com/docs/discovery-data?topic=discovery-data-about">Watson Discovery for IBM Cloud Pak for Data</a>” supports <a href="https://cloud.ibm.com/docs/discovery-data?topic=discovery-data-collections#configuredls">document-level security</a> for several data sources. You can review and potentially leverage that. If you have security requirements, concerns or are dealing with sensitive data, then you should reach out to your IBM contact to discuss.</p><h3>Capture Context</h3><p>The first thing you need to do is pass in and capture the logged-in user’s context in Watson Assistant. In my example, I will use the “role” of the user to determine the appropriate set of documents they should have access to and get answers from (in Discovery). Upon the start of each chat session in Watson Assistant, you can pass in context using the <a href="https://cloud.ibm.com/apidocs/assistant/assistant-v2?code=python#send-user-input-to-assistant">message API</a> using the context parameter.</p><p><em>(Example using </em><a href="https://github.com/watson-developer-cloud/python-sdk/blob/master/ibm_watson/assistant_v2.py"><em>Watson Python SDK</em></a><em>)</em></p><pre>import json<br>from ibm_watson import AssistantV2<br>from ibm_cloud_sdk_core.authenticators import IAMAuthenticator</pre><pre>authenticator = IAMAuthenticator(‘{apikey}’)<br>assistant = AssistantV2(<br> version=’2020–02–05&#39;,<br> authenticator = authenticator<br>)</pre><pre>assistant.set_service_url(https://medium.com/%E2%80%98%7Burl%7D%E2%80%99)</pre><pre>response=assistant.message(<br> assistant_id=’{assistant_id}’,<br> session_id=’{session_id}’,<br> input={<br> ‘message_type’: ‘text’,<br> ‘text’: ‘Hello’,<br> ‘options’: {<br> ‘return_context’: True<br> }<br> },<br><strong> context</strong>={<br> ‘global’: {<br> ‘system’: {<br> ‘user_id’: ‘my_user_id’<br> }<br> },<br> ‘skills’: {<br> ‘main skill’: {<br> ‘user_defined’: {<br><strong> ‘role’: ‘admin’</strong><br> }<br> }<br> }<br> }<br>).get_result()</pre><pre>print(json.dumps(response, indent=2))</pre><p>In my example above, I am leveraging a <a href="https://cloud.ibm.com/docs/services/assistant?topic=assistant-api-client-get-context">“skill-specific” context</a> to capture my user-defined entity in Watson Assistant named ‘role’. This context should be passed in from whatever front end the user is authenticating to access the assistant. You can learn more about handling context variables <a href="https://cloud.ibm.com/docs/services/assistant?topic=assistant-api-client-get-context">here</a>. Once you’ve got the context captured, you need to store it in the dialog node that will make the call out to Watson Discovery.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/355/1*CmTRbJI4vt5ioL0fN_fKmw.png" /><figcaption>Capturing context in your dialog node</figcaption></figure><p>The next step is to call Watson Discovery. Below are two different options you can consider when trying to implement document-level governance for Watson Discovery while using Watson Assistant as the front end.</p><h3><strong>Document-Level Filters</strong></h3><p>In this approach, you would ingest all of your documents into one Discovery <a href="https://cloud.ibm.com/docs/discovery?topic=discovery-getting-started#create-a-collection-tool">collection</a>. You can add a new parameter as <strong>metadata</strong> to your documents when ingesting them into Discovery. In this example, you would add a key/value pair with “role” as the key parameter and “admin” or “user” as the value, for example. The value would depend on the type of document and the intended audience. See the <a href="https://cloud.ibm.com/apidocs/discovery/discovery#add-a-document">“Add a Document” API</a> for details on how to do this.</p><p>Once you’ve ingested your document sets with the additional “role” metadata, you can use it as a filter in your query from the Watson Assistant dialog node. The user context captured in the step above is stored as the role and will be used as my filter. In this example, my dialog node will have a response type set to <a href="https://cloud.ibm.com/docs/assistant?topic=assistant-dialog-overview#dialog-overview-multimedia-add">search skill</a>, which calls Watson Discovery natively from Watson Assistant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/314/1*aWWiZRu3cYaeU4Ph0EbEKA.png" /><figcaption>Leveraging Discovery metadata as filters in your query</figcaption></figure><p>Below is a simplistic view of the result from the search skill. I opted to view the title of the document (which in my case is the file name) and the URL that leads to that document. This is configurable in the search skill itself.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/354/1*pOgrwDIIPsdCwF2ReEWRaA.png" /><figcaption>Search Skill response</figcaption></figure><p>This governs the answers presented to your users so each respective user role will get answers from the appropriate set of documents. The benefit to this approach is that it only uses one Discovery collection for all documents. That could also be a drawback if metadata tagging is not done efficiently. That is an additional task to complete.</p><h3>Using Webhooks</h3><p>In this approach, one Discovery collection is used for each respective set of documents (eg one for admins and one for standard users). Once ingested, the documents reside in different collections, different from the previous approach. You do not need to add additional metadata using this approach.</p><p>One of the response types in Watson Assistant is a <a href="https://cloud.ibm.com/docs/assistant?topic=assistant-dialog-webhooks">webhook</a> that allows you to make a POST request to an external API. In this approach, a simple program needs to be written. <a href="https://cloud.ibm.com/functions/">IBM Cloud Functions</a>, for example, is a good deployment option for the program. This program calls a Watson Discovery collection. <em>Explaining how to create a Cloud Function is out of scope for this post but you can view the </em><a href="https://cloud.ibm.com/functions"><em>documentation</em></a><em>.</em></p><p>The collection ID called in the Cloud Function will depend on the role context provided by Watson Assistant captured at the beginning of the chat session. The cloud function will use that role as business logic to decide which collection to query. Once the Cloud Function is created, the <a href="https://cloud.ibm.com/docs/assistant?topic=assistant-dialog-webhooks">webhook option </a>in Watson Assistant needs to be configured to call that endpoint.</p><p>Once that is done, configure the Assistant dialog. There are many different ways to set up the dialog node that will be invoking the webhook. Here is one example. The new node that will call the webhook can be added to the very bottom of your list of nodes, depending on the use case.</p><p>In that node, we need to catch every utterance that did not have a match in your previous dialog nodes:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/233/1*yZMlBrmgdOqW-Ij5yrvAdg.png" /></figure><p>Then check for the “role” entity to ensure you have that context:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/312/1*xk4swEqjC7UDnvS3Q2t4qQ.png" /></figure><p>Specify the parameters you will pass to the Cloud Function:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/643/1*Uo-X32LVj8sXY3sixlx3_A.png" /></figure><p>Then set the response conditions:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/567/1*PwtjvrH8448upt3QcTASWA.png" /></figure><p>Above, I added simple formatting to parse out exactly what I want to return from the answer that Watson Discovery provided to the Cloud Function. I opted to return the title of the document and the message body:</p><pre>&lt;b&gt;Title:&lt;/b&gt; “$webhook_result_1.title” &lt;br&gt; &lt;b&gt;Summary:&lt;/b&gt; “$webhook_result_1.message”</pre><p>Here is a simplistic view of an example response returned from Watson Discovery, leveraging the Cloud Function:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/389/1*sQc4yV4frDxkDbpZZM9evg.png" /></figure><p>The benefit of this approach is that you do not need to manually add metadata to your document sets. The other benefit is that the document sets will reside in their own collection, not mixed together. The additional work required in this approach would be to create the webhook but it is not difficult nor time consuming for someone who has experience developing a simple, lightweight application.</p><h3>Conclusion</h3><p>Both of these approaches can be leveraged to achieve document governance. Each has its benefits and drawbacks but ultimately deliver the same results. The approach you take will depend on your preference.</p><p><strong>Marc</strong> is<strong> </strong>the CTO for IBM Watson AI Strategic Partnerships. When not leading the technology vision and strategy for IBM Partnerships, Marc enjoys DJing, playing video games and wrestling.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9b6e1edd1e08" width="1" height="1" alt=""><hr><p><a href="https://medium.com/ibm-data-ai/document-governance-with-watson-discovery-and-watson-assistant-9b6e1edd1e08">Document Governance with Watson Discovery and Watson Assistant</a> was originally published in <a href="https://medium.com/ibm-data-ai">IBM Data Science in Practice</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Partnerships Have Become Critical To All Companies]]></title>
            <link>https://marcnehme.medium.com/why-partnerships-have-become-critical-to-all-companies-c59f5bea0823?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/c59f5bea0823</guid>
            <category><![CDATA[strategy]]></category>
            <category><![CDATA[collaboration]]></category>
            <category><![CDATA[partnerships]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[growth]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Thu, 30 Jan 2020 13:01:02 GMT</pubDate>
            <atom:updated>2020-07-30T20:44:54.115Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why Strategic Partnerships Have Become Critical To All Organizations</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RMW1dCkork-oiUBEDgQNKA.jpeg" /></figure><h3>Why are Strategic Partnerships Critical?</h3><p>We have entered a new era in the business world — the era of strategic partnerships. While partnerships have been happening for decades, there has been a boom over the last 5 years. Direct competitors have been partnering and collaborating for a mutual benefit. Alliances we thought we would never see, are happening. Synergy across different brands has become necessary. This trend will continue to grow in the coming years so you better hop on board while you can!</p><p>Successful companies look for continuous growth. They want to innovate and expand their business into different markets. Some will eventually hit a peak with their sales and/or user base and struggle to grow further. They still want to remain among the “cool” companies. So how can those companies continue to grow and stay “cool”? They form strategic partnerships with other successful companies to help scale their business.</p><h3>Best of Breed Partnerships</h3><p>An example of a strategic partnership in the technology industry is the IBM and Box partnership. This combines the “best of breed” in technology to be “better together”. Bringing <a href="https://www.ibm.com/watson">IBM Watson’s AI platform</a> to Box unlocks many growth opportunities for both sides. This benefits Box and IBM clients who are looking to improve their business processes with AI. Box is better with Watson AI, further differentiating themselves as a leader in the cloud content management space. Likewise, IBM is better with <a href="https://www.box.com/home">Box’s content management platform’s</a> added capabilities. This partnership allows Box and IBM to create powerful solutions powered by both leading technologies.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/353/1*3FOiy-TVL3gS4OaO2gpd7w.png" /></figure><h3>Partnerships in Competing Organizations</h3><p>There is more to partnerships. It’s important to note that to maximize the scaling of your platform, you need to appeal to a broader audience. This means partnering with one or more companies in your competing market. A great example of this is in the video game industry. Microsoft makes the <a href="https://www.xbox.com/en-US">Xbox console</a> and their major competitor, Sony, makes the <a href="https://www.playstation.com/en-us/">Playstation console</a>. These two companies have battled head to head in the video game market for over 15 years. Recently, they partnered to create new cloud-based solution and gaming experiences. They are combining the best of both platforms to create new offerings.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*v9WV7aq3YtzoVKA0FqKW9Q.jpeg" /><figcaption>Xbox &amp; Playstation controllers</figcaption></figure><p>Microsoft also has close relationships with video game juggernaut, <a href="https://www.nintendo.com/">Nintendo</a>. Nintendo has produced the most innovative video game platforms for the last 35+ years. The Nintendo Switch is their most recent console. At the end of the day, it is competing with Microsoft’s Xbox and Sony’s Playstation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*En7It8aNhGtJnTYACCONSg.jpeg" /><figcaption>Nintendo Switch console</figcaption></figure><p>Microsoft showed interest to collaborate with Nintendo. They wanted to expand their outreach with Nintendo consumers. Microsoft has released their own games, such as Minecraft, on the Nintendo platform. Microsoft also made their online gaming service available on the Nintendo Switch. Their goal is to cross-pollinate on the same game from two different platforms. It is expected that these two companies will collaborate more in the future.</p><p>Microsoft’s approach with Sony and Nintendo is a great example of an organization using partnerships to maximize their outreach in the market. Keep in mind the goal for these partnerships. Both sides should mutually benefit from the collaboration. If only one side is benefitting from the partnership, then reassessment is likely needed.</p><h3>Partnerships That Just Make Sense</h3><p>Another reason for strategic partnerships is to do what makes sense. Building all your required technology in-house is unrealistic for most organizations. It is expensive and resource intensive. Why reinvent the wheel when it’s already available? Assuming it costs less, why not partner with another vendor to have them handle the product development and support? It just makes sense.</p><p>An important factor for survival and growth in the business world is using the “best of breed” technology stack. Leveraging one company for all your technology will likely not meet all your needs. Your goals will require using many vendors to meet those needs across the board.Using that “best of breed” stack ensures you run your business most efficiently to enhance your products, services and value. Here are some examples of “best of breed” companies and their services:</p><ul><li><a href="https://www.ibm.com/cloud"><strong>IBM</strong></a> — provides world-class cloud hosting for your web applications along with a plethora of services. IBM Cloud provides leading Data and AI services such as <a href="https://www.ibm.com/cloud/watson-assistant/">Watson Assistant</a> and <a href="https://www.ibm.com/cloud/watson-studio">Watson Studio</a> to train and build machine learning models for AI-powered applications</li><li><a href="https://slack.com/"><strong>Slack</strong> </a>— provides a modern, collaborative platform to connect and communicate with your teams, optimizes your workflow, meetings and more.</li><li><a href="https://www.box.com/home"><strong>Box</strong></a> — provides enterprise cloud content management for your data as well as numerous tools and integrations to help you work faster</li><li><a href="https://www.salesforce.com/"><strong>Salesforce</strong></a> — provides a complete CRM platform to manage your customer information for sales opportunities, customer service cases and more.</li><li><a href="https://github.com/"><strong>GitHub</strong></a> — provides a complete platform to help developers collaborate together to write better code, manage issues, projects and more.</li></ul><p>So ask yourself…Is your organization partnering to increase outreach to scale their growth? Are you using the best of breed technology?</p><p><strong>Marc</strong> is<strong> </strong>the CTO for IBM Watson AI Strategic Partnerships. When not leading the technology vision and strategy for IBM Partnerships, Marc enjoys DJing, playing video games and wrestling.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c59f5bea0823" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Improve Your Sales & Product with this AI Pattern]]></title>
            <link>https://marcnehme.medium.com/generate-sales-leads-predict-winning-opps-drive-product-improvement-all-with-this-watson-pattern-245d0978f448?source=rss-5e91026c92ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/245d0978f448</guid>
            <category><![CDATA[unsupervised-learning]]></category>
            <category><![CDATA[tutorial]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[crm]]></category>
            <dc:creator><![CDATA[Marc Nehme]]></dc:creator>
            <pubDate>Thu, 12 Dec 2019 16:56:07 GMT</pubDate>
            <atom:updated>2020-07-30T20:45:10.560Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Czd2je7gp0z-WL5vCWHaHw.jpeg" /></figure><h3>Improve Your Sales &amp; Product with this Watson AI Pattern</h3><p>Many organizations struggle with both identifying and prioritizing what sales leads to pursue. Where do you start when you have a large stack of leads to go through? What do you when your leads have gone cold?</p><p>For Product Leaders, it&#39;s often a challenge to get a broad spectrum of feedback from their customers. How do they know where to focus next? How can they truly confirm their ROI? How can they improve the product with certainty to drive more revenue?</p><p>In this post, I’m going to share the details on a powerful pattern that my team and I got hands-on with. This solution pattern leverages many capabilities of the <a href="https://www.ibm.com/watson">IBM Watson Platform</a> to optimize your product and sales process. In this example, I will show Watson embedded in the CRM platform of Salesforce. Please note that you can implement this same pattern in other CRM platforms, such as Oracle, SAP, and others.</p><p>Below is a demonstration of this solution pattern in action:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FJ5SfgtmU_us%3Ffeature%3Doembed&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DJ5SfgtmU_us&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FJ5SfgtmU_us%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/a9dd382cfd1914d0000afc51dc6687c1/href">https://medium.com/media/a9dd382cfd1914d0000afc51dc6687c1/href</a></iframe><p>Here is an end-to-end view of the solution pattern workflow:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pkMo1XWYcj5K-vG1miakdA.png" /><figcaption>Watson-powered workflow</figcaption></figure><h3>Generating Leads</h3><p>Let’s start on the sales side — As a seller, you strive to pursue the highest quality leads, as quickly as possible. <a href="https://www.ibm.com/cloud/watson-discovery">Watson Discovery</a> can generate new leads from potentially untapped resources of millions of news articles, updated daily. This allows sellers to access leads on-demand.</p><p>Below, you can see Watson embedded in CRM, allowing the seller to specify exactly what types of leads they would like to pursue. A seller can provide any text as their search string and Watson will understand that natural language query. One can also specify company names to filter in/out, select the number of results to return, and choose an industry to target. All these options are made possible by Watson Discovery.</p><p>In this example, let us target the retail industry. I am looking for new businesses that are set to open in the city of San Francisco. Once I click on “Generate”, my query is sent to <a href="https://www.ibm.com/watson/services/discovery-news/">Watson Discovery News</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zzskZ86X-kdAuaWlgKV71w.png" /><figcaption>Lead Generation With Watson</figcaption></figure><p>I selected to return 5 leads, but in my result set below, you will see 4 leads. There is a feature that removes duplicate entries from the result set. You can also view the date the news article posted and a preview of the content. I will select the “Decathlon” lead and import it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZaZs3q_uO_Q5CDjTxqXswQ.png" /><figcaption>Lead Results</figcaption></figure><p>Once you click on import, the Watson news article is automatically converted into a Lead record in CRM. Watson will provide as much information from the lead that is available. Below are some basic data fields derived with Watson AI such as company name, industry and location information.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/826/1*3_zI6a65cO9fCTJ2Ku71FQ.png" /></figure><p>Watson provides other information that will be helpful to determine if this lead is worth pursuing, such as the article title and description. You can view more information on the lead source by clicking on the article URL provided by Watson.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xPiuO5_sayM7u864xgdUyA.png" /><figcaption>Data provided by Watson Discovery News</figcaption></figure><h3>Outcome Prediction</h3><p>Yet, there is still plenty of other data that can be analyzed related to this lead that can help Watson’s prediction. For example, you can contact this lead via phone or web survey to better assess their interests and needs. In this example, I have contacted the lead for more feedback. In their response, they provided the job family they belong to, their time frame to make a decision, and if they have an existing solution.</p><p>The key data point here is the “Key Objectives Input” field. This gives the customer an opportunity to state what they are looking to achieve. In this pattern, using the <a href="https://www.ibm.com/watson/services/natural-language-classifier/">Natural Language Classifier</a> service, Watson analyzes the unstructured text and classifies the intent to a structured field. In this case, security is this customer’s main goal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/687/1*2cTQcX5lxZt1YUibajIWOg.png" /><figcaption>Watson classifies unstructured text using NLC</figcaption></figure><p>In this pattern, we have trained a machine learning model using <a href="https://www.ibm.com/cloud/watson-studio">Watson Studio</a> and <a href="https://www.ibm.com/cloud/watson-studio/autoai">AutoAI</a>. We deployed the model in the <a href="https://www.ibm.com/cloud/machine-learning">Watson Machine Learning</a> service. The goal is to get a prediction on the likelihood of leads converting to won opportunities. Most of this data you have seen thus far is used by Watson to predict how likely this lead is to close as a win.</p><p>Now I can sort my leads by the highest probability to close as wins, thanks to Watson. As a seller, I would start at the top of the list to review and pursue further as an opportunity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NZlodD9tk1ACuu0pKDP6og.png" /><figcaption>Watson predicts the leads most likely to close</figcaption></figure><h3>Post-Opportunity Analysis</h3><p>Let’s fast forward the sales cycle where the lead was converted to an opportunity and closed as a win! There are always lessons learned and areas for improvement. After the opportunity is closed, sellers will obtain post-sales feedback. This can be an in-person conversation, a phone call, email, or other. That feedback contains lots of helpful information for various stakeholders in your organization.</p><p>Below is an example of some post-sales feedback received that was analyzed by Watson. Using the Natural Language Classifier service, Watson converted unstructured data into structured data. That data can be aggregated into reports and rolled up in dashboards. Here are three examples of the feedback obtained.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/738/1*s3jT8mO3TkblLmOXtPDnHA.png" /><figcaption>Post-sales feedback analyzed by Watson Natural Language Classifier</figcaption></figure><p><em>Sales Cycle Recommendation</em> — this guides Sales Leaders on improvements they can make to their sales process. In this example, the seller could use some coaching on better communicating with clients.</p><p><em>Product Gaps</em> — this helps Product Leaders understand what areas they need to focus on, to improve their product. In this example, the customer provided feedback that the product pricing is higher than competitors. This could be an indicator that the Product team needs to reassess their pricing strategy.</p><p><em>Top Features </em>— Product Leaders want to be aware of what is going well with their product to justify their ROI. It is also confirmation that the product is heading in the right direction.</p><p>So now, what do we do with all this data for each individual opportunity? How do you make it consumable to sales and product leaders? How does one learn from it? How does Watson continue to get better?</p><p>All of these closed opportunity records, both wins, and losses, are ingested into Watson Discovery for more AI analysis. Watson Discovery will enrich each opportunity record with its out of the box <a href="https://www.ibm.com/watson/services/natural-language-understanding/">Natural Language Understanding</a> model. It will also identify more patterns within the unstructured data.</p><h3>Performance Measurement</h3><p>Watson has done all the heavy lifting of analyzing large amounts of data. Those data points now reside in Watson Discovery and can be queried to create reports and dashboards to be shared with stakeholders. Below is a simple dashboard in Salesforce that shows some key areas for the organization to focus on.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eUBRhLLKE9l7F7ATjYReFQ.png" /><figcaption>Dashboard insights provided by IBM Watson</figcaption></figure><p>As a Sales Leader looking at the “Sales Feedback” report, I can clearly see that I need to provide some coaching to my team to improve their demeanor with customers. Watson’s insights show that this was the biggest factor that led to lost deals. For more detail, I can drill down into my reports to look at individual record data for specifics.</p><p>As a Product Manager looking at the “Product Gaps” report, I realize that I need to revisit my User Interface component. In contrast, looking at the “Top Features” report, Watson is showing me that my customers are enjoying the reporting component of my product. That’s a good confirmation of our ROI for that component.</p><h3>Continuous Learning</h3><p>Finally, this pattern implements continuous learning. Once opportunities are closed, lead and opportunity data is automatically fed back into Watson’s machine learning model. This ensures Watson continues to learn to get better. There is nothing required for an end-user to do in this mode of unsupervised learning. Full automation.</p><h3>Implementation</h3><p>Just a reminder that though this example was shown using Salesforce as the CRM platform, you could implement this pattern with other CRM platforms as well. Be sure to use our official <a href="https://github.com/watson-developer-cloud">IBM Watson SDKs</a> to reduce your development effort.</p><p>To learn more about leveraging the IBM Watson platform, contact your IBM sales representative.</p><p><strong>Marc</strong> is<strong> </strong>the CTO for IBM Watson AI Strategic Partnerships. When not leading the technology vision and strategy for IBM Partnerships, Marc enjoys DJing, playing video games and wrestling.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=245d0978f448" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>