Search Results

In our session at the Global Agility + Innovation Summit, a data scientist (that’s Pri ) and a product manager (that’s me) broke down what actually changes when you build AI products, what still holds true, and how to make smarter decisions along the way. Video Slides What Needs to Change When Building AI Products 1. Monitoring and Continuous Evaluation Are Critical AI features require ongoing evaluation after launch , unlike traditional software. Invest early in telemetry, monitoring, and feedback loops  to track performance, model drift, and degradation. V1 Trap : Teams often overestimate early performance and skip monitoring. Don’t. 2. Probabilistic Nature Requires a New Mindset AI outputs are not deterministic —they're influenced by context and often have multiple valid answers. Teams must grow comfortable with uncertainty  and open-ended results. Build fallback logic , success ranges, and user control into your features. 3. ML Teams Don’t Fit into 2-Week Sprints ML work often requires exploration, data gathering, and experimentation —things that don’t map cleanly to agile sprint cycles. Bring ML teams in early  for strategy and research—not just late-stage implementation. Long tail iteration  is a feature, not a bug. 4. MVPs Don’t Work the Same Way Traditional MVPs fall short for AI features; you can't always fake intelligence. Instead, prototype with Wizard-of-Oz tests, rules-based logic , or small models to validate before scaling. What Still Works When Building AI Products 1. You Still Need a Real Problem AI doesn’t make a weak idea better. Validated, specific user pain points  are non-negotiable. If it doesn’t pass the “so what?” test, don’t build it. 2. Prototyping and Feedback Loops Matter Just like any product, user feedback is gold . Test early and often, even before committing to building an actual model. 3. Cross-Functional Collaboration Is Key Success still depends on tight alignment between product, engineering, design, and now ML . ML should be a thought partner, not a service team . Red flag: “We’ll bring them in later.” What You Should Be Doing Start with Strategy What’s the business goal? How does AI help achieve it? Economic reality = fewer, smarter bets . Checklist: Clear business and customer value Focused use case Data feasibility Risk analysis Define Use Cases Before Writing Code Use the Objective → Use Case → Experience → AI Feasibility  framework. Validate with data , not just instinct. Checklist: Clear input/output Early data signals and KPIs What “good” looks like Engage ML Teams Thoughtfully Involve data scientists from ideation to implementation . Avoid the “handoff” model—treat ML as part of the core team. Checklist: Clear roles and collaboration points Feedback from ML team early and often Build a Feedback Loop Treat AI products like operational systems , not just code deployments. Monitor and iterate continuously  post-launch. Checklist: Tooling for feedback and performance monitoring Plan for long-term support and improvement Use Existing Tools Apply IA tools like card sorts, metadata taxonomies , etc., to support structure. Invest in data quality and governance  early. Checklist: Use proto-ontologies and personas Agile, test-driven development Resources Introduction to ML and AI - MFML Part 1   (Cassie Kozyrkov) 7 Reasons Why Most AI Projects Never Make It to Production  (Jan Van Looy) Your AI Product Needs Evals  ( Hamel Husain) LLM Evaluation: Everything You Need To Run, Benchmark LLM Evals   (Aparana Dhinkakaran and Ilya Reznik on Arize) All about LLM Evals (Christmas Carol on Medium) The definitive guide to AI / ML monitoring  (Mona Labs) The Real-World Harms of LLMs, Part 1: When LLMs Don’t Work as Expected  (Arthur AI) A Comprehensive Guide on How to Monitor Your Models in Production  (Neptune AI) Related Content from Hallway Studio The Business Leader's Guide to Writing AI Use Cases Why Ontologies Fast-Track AI Product Development Before You Bet Big on AI, Run Some Experiments Why I’m Not Data-First When It Comes to AI (Usually) The Business Leader's Guide to Writing AI Use Cases

What we consider valuable at work is shifting — fast. The skills, tasks, and roles that once defined success are being reshaped by automation, AI, and changing business goals. The people who will thrive in this new era are the ones willing to rethink how they deliver value, learn new skills, and let go of work that no longer matters as much even when that work once defined them. This is really hard. You worked hard over years to develop skills and it sucks to realize they aren’t valued in the market the way they once were. I’ve written before about this when discussing Immunity to Change.  Why AI Is Redefining Workplace Value Years ago, I was headed to a wedding in Long Island with my family. My mom, a proud New Yorker, was furious at us for using Google Maps. Why? Because she knew the “best” way to get anywhere — the right turns at the right times. That was how you showed value: you knew the way . But Google had live traffic data. It routed us over the GW Bridge and saved us 35 minutes. She was mad. We were on time. Value shifted. And we hadn’t told her. Fast forward years later, and she used AI to help write a tear inducing speech for her 50th wedding anniversary. We can all adapt. What AI Can’t Do (That Still Matters at Work) AI is great at: Summarizing conversations Drafting documentation Analyzing basic data But AI isn’t great at judgment . It doesn’t understand context, nuance, or trade-offs. Like when my GPS wanted me to take a risky left turn across four lanes at rush hour — a technically “faster” route that no sane person would actually choose. That’s where you come in. You add value by: Exercising judgment Asking better questions Seeing the big picture Making smart trade-offs How to Stay Valuable When AI Takes Over the Easy Stuff Let’s break down a few real examples to help you run some experiments to generate value in your unique situation. 1. Note-Taking and Summaries Old value:  A junior team member took notes and sent recaps. Now automated:  Tools like Fathom or Otter.ai do this instantly.  New value: Share your thoughts and proposals in the meeting so it’s captured in the transcript for others  Track action items and follow-ups Identify open questions and next steps Propose a decision or escalate a blocker 2. Status Reports Old value:  Manual summaries of projects, blockers, and plans. Now automated:  AI tools can pull logs and generate drafts. New value: Add analysis: What’s trending? Benchmark progress Highlight patterns or risks 3. Documentation Old value:  Writing step-by-step guides. Now automated:  AI can draft from meeting transcripts or demos. New value: Improve find-ability and usefulness Help clients integrate your product into their real workflow Track whether it’s actually helping anyone 4. Competitive Analysis Old value:  Feature-by-feature comparison in a spreadsheet. Now automated:  Perplexity or Claude can do that in seconds.  New value: Understand competitor strategy and positioning Identify gaps and whitespace Propose how your company can win — and draft messaging Questions That Will Help You Create More Value What am I doing that could be automated? What job is my stakeholder really trying to get done? How can I answer questions before they’re asked? What insights can I add that AI can’t? How can I take this task one step further or one level higher? Human Judgment Is the New Differentiator AI changes the game, but it doesn’t take you out of it. In fact, it makes your judgment  and insight  even more valuable. “A lot of the time when AI developers claim that AI can replace this or that job, they're doing so with a very narrow conception of what that job actually involves.” – Arvind Narayanan, Director at Princeton’s Center for Information Technology We’re not just doing different tasks. The definition of work itself  is changing. So don’t just do what you’ve always done. Ask what’s most valuable now . Final Thought: Let Go, Level Up You are not your task list. You are not your job title. You are your ability to adapt, to grow, and to deliver value where it matters most. Get after it and I’ll help.

If you’re the one making the decision and driving everythingIf teams are confused, endlessly debating and not making progress   The problem is a lack of direction I’ve known what the problem is for years and I’ve struggled to help people solve it. But you know me, I bounce back and try something new. I’m calling it the Direction Stack. It helps CEOs and executive teams turn strategy into a shared language teams can actually to make decisions and make progress  I’m not trying to change your strategy.  I’m trying to make sure your team gets it  and can use it . What is the Direction Stack? The Direction Stack is a communication and clarity tool for CEOs and executive teams. It helps you translate strategy into a shared language your teams can use in: Planning Roadmapping Prioritization Tradeoffs Hiring The goal isn’t to impress investors (but it will likely help you pitch better)It’s to make sure your team can explain what matters, why it matters, and how to act on it What’s in the Stack? 1. Direction One Liner What are we building and why does it matter right now? This is your team’s quick reference point—a clear, present-tense statement that connects daily work to purpose. Everyone should be able to say it out loud without a slide. A good Direction One Liner is: Focused on what’s happening now Ambitious but concrete Easy to remember and use in decisions Example: “Developers are using our product. Now we’re building features that engage security teams, so we can land bigger deals and expand in the enterprise.” Why it matters: If people can’t explain what they’re doing and why it matters, alignment drifts and decisions get slower. 2. Strategic Focus Areas What 3 bets are we making? Make the case for the priorities that will drive progress over the next 3 to 6 months. Format: We’re focused on [What] , because [this is how it moves us in the direction we want] . Examples: We’re improving onboarding, because activation in the first five minutes drives long-term retention. We’re launching a European data center, because regulatory requirements are blocking deals. We’re shifting resources to customer success, because week 4 churn is pushing down LTV. Why it matters: Without the “why,” priorities feel like busywork. That leads to low morale, slower delivery, and less innovation. 3. Manager Tools How do we keep everyone aligned in meetings, updates, and decisions? Managers are your multipliers. Equip them with tools that reinforce priorities and guide decisions in the moment. Decision Filters Short, sticky rules that speed up aligned decisions: Client satisfaction over internal convenience Structured offers over custom SOWs If it’s a maybe, it’s a no Weekly Team Update Format: This week we’re focused on: [Tactical effort] It supports: [Strategic Focus Area] It matters because: [Why the strategy exists] We’re tracking progress with: [Metric or indicator] Prioritization & Tradeoff Questions: When teams are debating what to build, cut, or delay, ask: Does this connect to a strategic focus area? What’s the “why” behind this work? Which decision filter applies? How the Stack Operates The CEO owns the Stack since It’s your translation layer between vision and execution. It doesn’t replace your strategic planning or performance managements. It’s an add on that contextualizes and makes it useful. Create a new one every time your strategy shifts—or when you notice teams aren’t moving the way you hoped. Everyone, from execs to ICs, should understand what you’re doing, why it matters, and how to talk about it. Use the stack in planning, standups, retros, and tradeoff conversations. That’s how you turn clarity into action. How You Know It’s Working Anyone on your team can describe the strategy and why it matters Prioritization feels faster, because direction is clear Fewer things get “stuck” waiting for you Teams make better decisions, on their own If your team can’t  tell you the strategy—or apply it without you—it’s not clear enough. Why I’m Sharing This This isn’t a product. It’s a practice. Something you can test, refine, and evolve with your team. If you’re trying to move faster, align better, and reduce friction, try building your own Direction Stack. I’d love your feedback: What resonates? What’s still unclear? What would you change?

Many teams aren’t lacking strategy, they’re struggling with Direction Disconnect : the gap between big-picture plans and everyday decisions. You can hang your vision on a wall or drop it into your email signature. But if your team can’t apply it today, they’ll fall back on habit, guesswork, or whoever speaks the loudest in the room. That’s why we created the Direction Stack , a practical set of tools to make strategy repeatable, usable, and actionable at every level of your organization. The Problem With Vision and Mission Statements Companies love a good vision or mission statement  Canva: Empowering the world to design LinkedIn: Create economic opportunity for every member of the global workforce TED: Spread ideas Databricks: Simplify and democratize data and AI, helping teams solve the world’s toughest problems These statements are powerful. They guide long-term direction and big bets. But they’re too abstract to help your team decide what to build this sprint, who to prioritize, or which tradeoff to make. That’s where the Direction One Liner comes in. What Is a Direction One Liner? It’s your current, tactical version  of the vision. It helps your team: Know who they’re helping Understand what they’re doing and why See how today’s work ties into tomorrow’s goals It’s present-tense  for a reason. Your strategy evolves. This helps your team make smart decisions at the moment. Then you update it when we’re ready to get to the next step  How to Write One Use this format: We’re helping [group] [do this], so they can [get result] and we move closer to [long-term goal]. Or, when priorities shift: In the past, we focused on [X]. Now we’re helping [Y], because [change], and that sets us up for [future goal]. Real-World Examples We’re helping insurance companies get accurate, up-to-date data on the value of customers’ lost property—so they can settle claims faster and with more confidence. That also grows our used property data business. We’ve built a strong network of freight carriers, giving us data on capacity and routes. Now we’re onboarding shippers—so we can optimize matching, improve pricing models, and deliver better outcomes for both sides. We recently acquired a company with valuable customer data. We’re integrating it to improve prediction accuracy and open new cross-sell opportunities with existing products. Why This Works A good Direction One Liner: Anchors your team in what matters right now Connects strategy to execution Improves decision-making Reduces misalignment and drift Makes vision feel alive, not abstract Bonus: it helps you spot  drift , when different teams say different things, it’s time to realign. How to Use It Write one every quarter  or whenever priorities shift Repeat it  in sprint kickoffs, team meetings, and async updates Pressure-test it : Could a new hire repeat this after hearing it once? If not, make it shorter and clearer. Bottom Line You don’t need a new vision. You need a way to operationalize the one you have . The Direction One Liner  helps your team move faster, make better decisions, and stay aligned as you scale. Want to see how this fits into the full Direction Stack?  → Read the post on fixing the Direction Disconnect Need help writing yours?  → Let’s work on it together

If you are early in your career, navigating org politics, or just trying to make a dent in a big system, this playbook is for you. These tools were built from trial, error and by learning from a lot of smart people.  Understand the people and system (even the weird and irrational stuff) When people join the workforce they are often frustrated. They’ve developed an idea of the way things should work from their parents, teachers, friends and social media, what they discover is a workplace that doesn’t make sense. We like to think that we make decisions rationally but we don’t.   We have many biases . We can be tricked by   behavioral science . We don’t use words the same way . The way we make decisions and communicate based on our experience and that is unique to us. So we understand it but no one else does. Organizations are made up of people, so no wonder they adopt our quirks and irrational behavior. We understand and accept your clients, stakeholders, teammates and executives as individuals. We can’t make our clients or stakeholders be the people we want them to be. They don’t speak our language, they are motivated by other things and their experience is their own. Instead of getting frustrated we need to meet them where they are. What seems weird and irrational to you might make lots of sense when you start to consider: Motivation:  Jennifer Goldman-Wetzler says we have   ideal, real and taboo values . Some values are easy to discover, others are hiding beneath the surface. Companies have the same hidden values. Their mission statement may say one thing, but you’ll find other values drive decisions. Constraints:  Barriers real or imagined exist in any organization. If you are new to the organization you may not understand what these are and how people are choosing which battles to fight. Communication:  Social Psychologist Heidi Grant Halvorson ‘s research shows how people interpret our words and actions. She’s found stereotypes and past experience can create miscommunications. Give people the benefit of the doubt and work to correct miscommunications. When you accept this then you can get to work learning about the people and what they care about, how they see success, how to motivate them and what will set them off. You’ll learn who you need to approach, how and in what order. Most importantly you can accept that change without power is a ground campaign. You win it by inches and your only tool is influence. If you think saying something smart is all you need to make your idea a reality, you will have a career filled with disappointment and frustration. You need help, money, permission and support and getting it isn’t easy. Prepare for the worst  So many times my conversations have gone sideways or I ran right into a brick wall. I’ve learned to prepare myself for the worst, have a plan and be ready for anything. If you’ve taken the time to plan you’re calmer and more confident and the person you’re talking to will see that. I've learned from influential people to whiteboard and plan conversations. We’d think about our audience and the story they needed to hear. We would make sure we had the answers to their questions or objections at the ready. I generally find worksheets silly but my teams have benefited from having templates for common hard conversations and doing some planning. You may think this is too much but it’s way better to be over than under prepared. Let’s say you have a great idea for a solution to a problem that’s been plaguing your team and you can’t imagine why anyone wouldn’t immediately say yes. Except when you explain your idea to your team lead the conversation goes wrong and your idea goes no where. Presenting a new idea, pushing back on a bad one, asking for something, delivering bad news, raising risks are all common conversations that can go wrong. Through trial and a lot of error I’ve learned that a little preparation helps things go more smoothly. I’ve created a series of worksheets for my team so they can self-serve if I’m not around. Don’t just influence, make an ask  If you want someone to do something (give permission, take an action, invest money) you need to ask them to do it. I’ve seen many people present ideas and but not ask for action. Don’t assume that your audience can read your mind. Be prepared to answer questions and objectives  Consider the perspective of the person or people you are talking to and list all of the questions, concerns or motivations they have. Then start thinking of ways to answer so you are prepared if it comes up. Always set the table for the conversation  I’m impatient, I always want to dive right into discussion. My friend Cip is the master of the setup, he makes sure that the person(s) he’s speaking with know the background, objectives and what we want to accomplish in the conversation. Clients, stakeholders and especially executives are very busy people, quickly remind them of what we’re doing and where we are in the process so they can make a decision. From there you can describe what’s happening and why it is a problem or opportunity. For example, “We are trying to do these stories, but we’re blocked because an api isn’t ready. We’re at risk not having the feature for the March release when the sales team needs it for the conference” Practice or role play (yes really) It can be incredibly awkward to practice or role play a conversation. But going through things with a friend will help you think through things, refine your point of view and give you more confidence. You may worry that you will become too rehearsed or robotic, I seriously doubt that will happen. Manage yourself According to Patrick Lencioni’s   Getting Naked , our fear causes us to withhold ideas, hide mistakes and edit to save face. What motivates you? drives you? annoys you? infuriates you? scares you? These are the things that are driving your elephant whether you know it or not. The more you accept this, the more successful you will be in emphasizing, managing emotions and ultimately creating change. When someone doesn’t like our brilliant idea it's easy to get frustrated and defensive but that’s only going to hurt your cause. If your ego is more important than your cause, you will not be successful. In   Getting Past No , William Ury argues that when things are hard our natural reactions are to strike back, give in or break off. Instead of reacting, we should “Go to the Balcony” to step back from our emotional reaction and try to get a broader perspective as if we stepped off the stage and moved to better vantage point. Ury also recommends that we know our hot buttons, those things that will make us angry or defensive. Read the room or person Sometimes we get so wrapped up in making our point or achieving our goal, we don’t focus on others. Are they interested, indifferent, annoyed, confused, jarred? Pay attention and make the effort to connect, be willing to say “you look confused” or “you seem concerned” so you can deal with any issue quickly. Hostage negotiators call this labelling.   Be on the lookout for what   Bob Rogers   calls an “approach response.” I hate split pea soup with a passion and just thinking of it on the stove in my parent’s house makes me ill. My approach response to pea soup is yuck.  In sales conversations or customer interviews I like to have a designated watcher who is focused on all the non-verbals and reactions so that we can learn from those as much as what they said. AI transcription tools can’t do this…yet.  Look for invisible drivers  There is often something that is spooking the elephant that isn’t on the surface. I’ve written before about the immunity to change  and how it holds progress back.  My mom worked for the federal government and as big government does they made a decision to ban space heaters because they were causing computer damage and increased fire risk. One of her employees kept resisting, it was getting so bad that a formal rebuke was the next step. My mom called the employee into her office and asked what was going on. The woman’s answer? “I’m cold.” When my mom decides something needs to get handled, it does and In 10 minutes the building contractor was fixing the heating. It’s always stuck with me as a good example of what’s driving our emotional reaction and how important it is to investigate it. Manage the emotions If there are seven people involved in something, you have eight sets of emotions to manage. Their  emotions and your own.  In their terrific book   Switch  the Heath brothers talk about the Rider and Elephant. “(psychologist Jonathan Haidt) says that our emotional side is an Elephant and our rational side is its Rider. Perched atop the Elephant, the Rider holds the reins and seems to be the leader. But the Rider’s control is precarious because the Rider is so small relative to the Elephant. Anytime the six-ton Elephant and the Rider disagree about which direction to go, the Rider is going to lose." We have to understand and manage these elephants. Change can make an elephant antsy, you need to see that anxiety and deal with it directly. It can often take time and a lot of touch points to get the rider and the elephant where you want them. If you spook the elephant, it’s going to take a calming and firm hand to get it to calm down and it may take several attempts to make the change. Don’t get discouraged, just be creative and keep working until you get there. Make awkward OK Sometimes there is no way around it, you have to get into the awkward stuff. The best way is to admit it’s going to get awkward up front. Former hostage negotiator Chris Voss suggests you use an accusation audit  to take the sting out of something.  Beware the dark side  You can use your powers for ill purposes. You can get so good at playing the system that you don’t stop and fix the system. That you influence when you should lead. I’ve learned this the hard way.

Many CEOs miss the early days of their company. When a small group of people could do amazing things. Everyone was on the same page. You moved quickly. You could change direction, learn, and feel like you were making progress every day. But now? It’s harder. You’ve got more people. More meetings. More decisions. And somehow, everything takes longer than it should. You have to be involved in things you don’t want to be involved in, but if you’re not, nothing moves. You look at what teams are working on and wonder: Why that? Not this?  You don’t feel like people understand where you need them to go. You don’t want to micromanage, but if you don’t step in, things drift. You pull everyone into a big alignment meeting. You lay it all out. And a week later? It’s like it never happened. This isn’t just your company. It happens often when a company goes through an inflection point  I’ve Seen This Problem for Years I’ve seen this as an employee. As a consultant. As a vendor. And here’s the thing I’ve learned: Most people in an organization don’t really know what they’re supposed to be doing. And they definitely don’t know why. So instead of forward motion, you get: Debates about what matters Conflicting opinions about goals Projects that get started, stalled, or shelved And that leads to real damage. Morale tanks as people feel jerked around.Progress slows because no one knows what to push. Trust erodes as goals shift in unpredictable ways  Leaders feel it too. The pressure. The frustration. I have tried to solve this problem and haven’t quite gotten there  I’ve written blog posts. I’ve given talks ( including a TEDx ). I don’t think I’ve helped people solve it as much as I want to. So I’m going to try again. Because I know what it feels like to be on a small team doing great things. It’s hard. It’s fun. It’s meaningful. You build something together and your bonds are strong. . That’s what I want for leaders and teams. But right now, too many of them don’t have that. The Direction Disconnect Here’s how you know this problem is showing up: Nothing moves without you. Teams are going in the wrong direction—or not moving at all. Your big meetings feel like you’re hitting reset, over and over. Execution slows. Morale dips. Friction rises. You might think this is a performance problem. Or a process problem. And yes, you might have some of that too. But if your team doesn’t have clear direction  and a connection to purpose , even your best people won’t succeed. You can’t process your way out of a direction problem.And replacing teams is time-consuming, expensive, and painful. Fixing direction? That’s easier than you think.It can happen in a day and you’ll know almost immediately if it’s working. How direction gets lost In the early days, everyone hears the strategy straight from the top. They know the mission. They get the vibe. But as you grow, that signal starts to break down. It’s like a radio: strong if you’re in range. Garbled—and eventually silent—the further you get. What your company needs is a signal booster. You need other leaders who aren’t just repeating the words—they’re translating them. Making them relevant for their part of the business. So every team, at every level, knows: Where we’re going What matters right now How to decide what to do next Imagine This Imagine sitting at a table with your CEO. No phones. No laptops. They lay out the big vision. The priorities. The decisions we need to make. And for the first time in a while, you think:  “I get it. I know why I’m here. I know what we’re building. I know how we will win.” That feeling? That energy? It’s powerful. But most people in growing companies never get to feel it. They never get that level of clarity or focus. Sometimes, it’s because the CEO doesn’t communicate clearly. Other times, it’s because the message falls apart after the meeting. But what if it didn’t? What if your message was crystal clear —and repeatable? What if other leaders could share it with confidence , adapting it for their teams while keeping the core intact? What if direction was part regular planning, prioritization, tradeoffs? That’s not a pipe dream. It’s doable. Direction Can Be Designed What it takes is: A clear message Captured in a simple, reusable format With tools for leaders to apply it in real decisions Without that, your teams will just keep reacting to the last loudest thing.And your strategy will sit in a deck somewhere while everyone else makes it up as they go. I’ve seen this gap at every level of growth. I’ve seen what happens when it’s not fixed. And I believe we can do better. So What Would That Look Like? What would it look like if your teams had a stronger connection to the vision? If every leader had a toolkit for: Explaining where we’re going Making priorities clear Helping teams make smart decisions day to day I think it would look like progress .It would feel like teamwork .And I think everyone would have a hell of a lot more fun . So I’m going to try and help. I’ll introduce some tools to do just that. You can tell me what works, what doesn’t, and what you need next. Let’s build teams and companies we want to be part of. And let’s finally fix the direction disconnect.

Scaling a business and implementing an AI strategy successfully requires more than technical skills. It demands high-performing product teams that are resilient, creative, focused, and deeply empathetic. Whether you're defining your AI roadmap for business or moving from startup to scale-up, these five team traits will determine if you thrive — or stall out. Here’s how to hire and build the kind of team that will drive your AI implementation strategy and help you scale with confidence. Hustle: Finding a way  Hustle is all about creativity and determination. It’s finding another way when the conventional path is blocked. It’s knowing when to charm, when to needle, when to escalate and when to throw down. It’s not getting stuck or frustrated when things aren’t going the way you wanted them to. In AI projects, hustle means finding creative ways to source, clean, and label data — often when it doesn’t exist neatly in a database. It means navigating a fast-moving regulatory landscape, overcoming skepticism about AI value, and finding early pilot partners willing to take a risk with you. So when you’re building your team look for people who: Are resilient to change and challenges Build relationships before they need them Understand and navigate bureaucracy Can come up with crafty and unique solutions to problems Aren’t afraid to talk to just about anyone Discipline: Staying focused amidst AI hype  A new idea is exciting and addictive. You can see the possibilities everywhere. Every new prospect or client I encounter is deeply passionate about their new idea. The trap is that you try to go in too many directions at once and end up getting nowhere. Innovative teams have the discipline to narrow the world of possibilities down to a single thing to focus on. Their discipline allows them to test, learn and iterate without getting off track. They focus on today’s problem of finding customers and meeting business objectives instead of tomorrow’s problems like scale. AI teams face the constant temptation to chase every new model or trend. Disciplined teams stay laser-focused on solving real business problems and defining measurable outcomes. Discipline helps you resist scope creep when every stakeholder wants their own chatbot or predictor. That’s not to say that innovative teams never get off track, because they do. What’s different is that someone notices, they point it out and the team refocuses their effort. I have a favorite wrangler. She has an amazing ability to help people focus their efforts on what to do right now and what to save for tomorrow based on business objectives.  I’ve written about one of our favorite exercises for this . It also doesn’t take long before clients adopt her way of thinking and are able to move quickly. When your building your team look for: The ability to create a framework that helps people focus in and make the hard tradeoffs People who know the difference between today’s problems and tomorrow’s problem and communicate effectively People who can steer the group back to the right direction if they get off track Inquisitiveness: Curiosity Powers AI Innovation In our office there is always a lot of discussion in person and over email about new technologies and trends. It’s this passion and interest in different verticals, behaviors, trends and technologies that that fuels innovation in a couple of important ways. Research shows several benefits for curiosity on teams  Increased Engagement and Motivation (Kashdan et al., 2018) Improved Problem-Solving and Creativity (Reio & Wiswell, 2006) Better Adaptability to Change (Vogl et al., 2019) Higher Performance (Litman, 2008; Reio et al., 2006) Knowledge Sharing and Collaboration (Reio & Wiswell, 2006) Broad Inquisitiveness  A lot of innovation comes from taking a concept from one space and moving it to another space. The   Wright brothers  experience in building bicycles led them to take a different approach than their competitors and contributed to them creating the world’s first successful airplane. My experience working with different types of companies and non-profits gives me a wide array of inspiration to draw from. Some of the best AI innovations come from borrowing ideas across disciplines: using anomaly detection techniques from cybersecurity to catch fraud in banking, or applying recommendation engine logic to employee training platforms. Teams curious across domains find unexpected wins. Sometimes I encounter folks looking for partners that only want to work with experts in their industry. The good news is that you will get a partner that knows your industry but you might get the same thinking and ideas as everyone else. If you want an innovative team, look for people who bring a diversity of experience and a passion for continuous learning. Deep inquisitiveness My favorite thing about working for a museum was getting to dive into a new subject every couple of months. I learned about history, international politics, photography, technology and more. Then I’d take what I learned about and find a way to make it engaging for a more general audience. Whatever the topic is, innovative teams dive in and quickly understand the business, customers and competitors. But they manage to keep an outside perspective and don’t get boxed in. When your building your team look for : Diversity of experience People who are always reading and sharing new ideas from different areas People who can dive into a new area and bring fresh perspective A resistance to conventional or internally-focused ideas Empathy: Essential for Human-Centered AI  Marketing professors Kelly Herd (University of Connecticut) and Ravi Mehta (University of Illinois)  ran a study with more than 200 adults to come up with potato chip ideas for pregnant women . Half was given the assignment directly and the other half were encouraged to imagine how the pregant women would feel while eating it. "Eliciting empathy has inherent value in maximizing creativity," Herd says. "We've shown that empathy can change the way in which you think…imagining how someone else would feel, can have a huge impact on creativity in general." I had the privilege of doing a pro-bono strategy workshop for   St. Jude  and toured the hospital. It’s an amazing place. When a patient arrives at St. Jude they are transported in a red wagon instead of a wheelchair, at registration the desks are low enough patients can see and interact with the staff, every space is a colorful one and in one hallway are pictures of former patients one of whom is a PhD researcher at St. Jude. Empathy helps them see the hospital experience through the eyes of patients, parents and siblings. The team understands pain, needs, behaviors, constraints, goals and values of the people they serve. Because of this they can create meaningful and impactful experiences. Will a customer trust an AI decision? Will they understand how it was made? Teams need to walk in the user's shoes to ensure AI augments instead of replaces human judgment." Innovative teams solve business and customer problems. These problems have an emotional and experiential component that must be understood to get the right solution. That solution needs to provide the right solution for the customer and business. When you're building your team, make sure there are a few people who: Actively seek understanding of the customer, business and stakeholders Watch facial expressions Put themselves in the shoes of others Advocate for the little details that make an experience better Make decisions based not on their own preferences but on how they think that decision will affect someone else Care not just about the customer but the business and how staff will be affected by the product When something is missed or wrong (and something always is) they are quick to spot the problem and fix it Humility and Resilience: Critical for Sustainable AI and Scaling Success I told a   CEO  I was going to get things wrong. I delivered on that promise and will continue to. I’m one of those people that needs to learn by getting bruises. I don’t particularly enjoy getting things wrong but I would rather have a small fail than an epic one. I’ve had to eat my words and my ideas. I’ve crashed and burned in front of seven year-olds, been wrong about pain points, pitched a value proposition that had no value to the users and have had teachers look at me like I’m nuts. I invite and expect this. In AI, even your most brilliant model will eventually degrade as the world changes around it — a phenomenon known as 'model drift.' Teams that succeed aren't the ones who get it perfect on the first try, but the ones who monitor, adapt, and rebuild without ego. Humility is accepting that you don’t have all the answers and being willing to engage your team, stakeholders and most importantly customers.  Resilience is picking yourself up when your tests fail. Having both will keep your team headed in the direction of a successful product. When your building your team look for people who: Are willing to take risks and make mistakes Engage others in brainstorming Value learning from mistakes Openly discuss challenges they’ve faced and how they overcame them or didn’t Conclusion: The Future Belongs to Adaptable Teams The world of AI and business scaling is changing fast. Building high-performance product teams isn’t optional, it's your edge. By focusing on hustle, discipline, curiosity, empathy, humility, and resilience, you’ll create a team that can deliver exceptional products, navigate your AI roadmap for business, and scale smarter and faster than the competition. Scaling with AI means embracing uncertainty, experimenting wisely, and staying relentlessly human-centered. The teams who do will shape the future." Note: This is an update to posts in my archive Image: A fun watercolor experiment my daughter and I did together

Thank you AllThingsOpen.ai for having me speak at this great event. The title of my talk—"You Don’t Need an AI Strategy, But You Do Need to Be Strategic About AI"—might sound provocative, but that’s intentional. Below you'll find the slides, citations, resources and the video. Too many companies are getting caught in "strategy overload," layering AI strategies on top of data strategies, marketing strategies, and tech strategies. The result? Confusion, competing priorities, and a whole lot of meetings with no clear direction. So, let’s reset. This talk is about how to integrate AI into your business strategy—without letting it become a distraction. Key Takeaways from the AI Strategy Talk Strategy Overload Is Real Raise your hand if your company has an AI strategy. Now keep it up if it actually aligns with your business strategy. If your AI efforts don’t support your overall business goals, you’re just chasing hype. AI Isn’t Magic—It’s a Tool AI can automate, analyze, personalize, and recommend—but not all AI projects are good investments. A bad AI use case is one that could be solved with traditional software or doesn’t have the right data to train a model​. Thinking Like a Strategist Instead of asking, “Do we need an AI strategy?” use these questions from Ethan Mollick . What’s valuable today that might become obsolete? What was impossible before but is now feasible? What can we democratize or personalize with AI?​ A strong AI use case checks these boxes:​  ✅ Clear value for both business and customers.  ✅ Uses unique data that gives you an advantage.  ✅ Has a baseline for comparison.  ✅ AI is actually necessary—not just for the sake of AI. Bad AI use cases?   ❌ Could be done with traditional software.  ❌ No clear patterns in the data.  ❌ Requires absolute precision (e.g., medical diagnosis).  ❌ The cost of implementation outweighs the benefits. Videos Slides AI Use Case Worksheet Citations Advice for finding AI use cases (Cassie Kozyrkov) AI Transformation Playbook (Andrew Ng) Working Backwards: Insights, Stories, and Secrets from Inside Amazon  (Colin Bryar and Bill Car) The Flywheel Effect (Jim Collins) Tech at Work: What GenAI Means for Companies Right Now (HBR IdeaCast featuring Ethan Mollick) Resources Using the term ‘artificial intelligence’ in product descriptions reduces purchase intentions (WSU Insider) Introduction to ML and AI - MFML Part 1 (Cassie Kozyrkov) 7 Reasons Why Most AI Projects Never Make It to Production (Jan Van Looy) Related Content from Hallway Studio The Business Leader's Guide to Writing AI Use Cases Why Ontologies Fast-Track AI Product Development Before You Bet Big on AI, Run Some Experiments Why I’m Not Data-First When It Comes to AI (Usually) The Business Leader's Guide to Writing AI Use Cases

You’re Not Being Repetitive Enough I can’t tell you how many times I’ve heard a CEO say, “But I already said that.” And here’s the thing: you may have said it—but that doesn’t mean your team heard  it, understood  it, or remembered  it. Leadership isn’t about one moment of clarity—it’s about consistent communication. If you want your team to align around your vision, internalize the strategy, and make the right decisions day to day, you need to repeat your story. And not just you. Your direct reports, their managers, and every leader in the org must be telling the same story in different rooms. That’s how shared direction actually happens. Repetition Is a Strategic Lever A recent study found that leaders are nearly 10 times more likely to be criticized for under-communicating than overcommunicating.  Nearly 75% of employees rated their managers as under-communicators , falling short of expectations for task-relevant communication. Even worse? Under-communicators are seen as less empathic and less qualified.  On a 1–5 scale: Under-communicating leaders scored 2.93  on empathy. Over-communicating leaders scored 3.64 . Leadership ability was rated at 2.20  for under-communicators vs. 3.60  for those who said more. You’re not going to annoy your team by repeating the same message—you’re going to build trust and credibility. When leaders communicate frequently and clearly, they’re perceived as more capable and more caring. Repetition Builds Fluency, Truth, and Action Professor Adam Galinsky from Columbia Business School puts it this way: “You want to repeat it over and over and over again. It creates a sense of fluency, which increases understanding and truth, value and validity.” In other words, repetition doesn’t just help people remember your message—it makes your message feel more true.  That’s a powerful insight for any leader trying to build belief and alignment. And if repeating yourself feels awkward? Good. “We can train ourselves to repeat that message over and over and over again,” Galinsky says. “It’s just doing any type of exercise, like working out. I work out with a trainer and the first time I do an exercise, it feels totally unnatural to me and sometimes impossible, even that day. But over time, it becomes more possible the more that we work at it.” You can train for repetition. You can build that muscle. And once you do, your message becomes part of the culture—not just a line in a slide deck. Practical Ways to Reinforce Your Story You don’t need to sound like a broken record. You need to sound like a leader who knows what matters. Here’s how to do it well: Say the same thing, in different ways Tell a story. Use a metaphor. Share a stat. Reference a customer. Show a visual. Bring it to life from different angles so it sticks in different minds. Anchor every message to your core story Project updates? Tie them back to the strategy. 1:1s? Connect the work to the mission. All-hands? Roadmapping? Repeat your values. Make your story the thread that ties it all together. Make repetition a team sport This isn’t just your job. Ask your VPs and directors to carry the message. Build a communication cascade where everyone knows the story and can share it consistently with their teams. Watch for echo One of my favorite moments when I was a consultant was when I heard someone else using my words, framing a decision with our shared story. That’s the signal that the message is working. Conclusion: Your Message Isn’t Real Until It’s Repeated You may think you’ve said it enough. Your team is still figuring out what it means. If you care about being understood—really understood— say it again . Then teach your leaders to do the same. Build rituals around the message. Embed it in your stories. Repeat it until it becomes second nature. That’s not redundancy. That’s leadership. And the best part? The more you do it, the easier it becomes. Just like working out. Ready to build that muscle? Say it again. And again. And again.

Behind every “overnight success” is years of work in the shadows—and the same is true for artificial intelligence. AI chat bots may have exploded and captured public attention but scientists, mathematicians, and philosophers have been working on this since before the 1950s according to Harvard.  For a long time we’ve used AI/ML for limited applications and now we see the opportunity to create efficiencies, make smarter decisions, supercharge employee capability, create new products, and have even more cat videos. What’s changed is that these tools are out of the lab and available world wide to regular people ready or not. Even CTOs are having a hard time keeping up with all the rapid changes and advancements in AI while trying to decide how to use AI to create and grow businesses. To help, here are six things the experts want you to know. You may be able to do it, but no one may be able to afford it AI may provide the best solution to your problem but the price tag to go from an idea to something working in production may be hundreds of thousands of dollars according to CTO and technology advisors I’ve interviewed . While this is likely to go down in price in the near future it can be challenging to justify the expense since most organizations are focused on keeping costs down. Most of the work is cleanup Experts estimated about 80% of the job is cleaning data and 20% is training and testing. It is a significant effort to do this cleaning and if you discover you don’t have the right data or enough data you’ll have to start over again. Sometimes teams will use data from a proxy group instead of their desired group and find out that they are off in their projections and have to rework everything. Speed and accuracy really matter How will data be served to the model? How can we fast serve it? A Fortune 500 company requires a response time to under 100 milliseconds and will keep shaving the time down to get under the bar. Speed isn’t the only important metric, the same company requires 95% accuracy before an AI-powered feature can be released. Understanding your audience is important in setting your success metrics.,While an audience of professionals may expect perfection where a more casual audience is fine with getting a general idea of what’s happening. Nobel winning psychologist Daniel Kahemen found that while we forgive humans for making mistakes we expect very high performance from machines or AI . People are always the hard part Technology makes it easier and faster and AI is no different, but companies are run by people. Leadership teams will need to decide together what their AI strategy is, what investments to make, what rules or governance to put in place, and how to lead their teams. These are going to be challenging conversations that will require vision, strategy and people skills to navigate. Explainability is a work in progress Some B2B companies are using traditional forecasting and modeling instead of AI because an audience of experienced professionals is unlikely to put their trust in something they don’t understand or explain when pressed by leadership or clients. Additionally there is a risks models are not complying with regulations like fair housing or fair credit. Double check, always An LLM is the loud person in the bar who is always confident, not always right, and shouldn’t be trusted with the keys. They may be able to pass the bar but make up cases for a legal filing . So just double check. Image: This is an old mixed media piece I did by cutting up the strips and putting them over a watercolor painting.

Years ago, my washing machine broke, and I had a repair person come by. By the time we got to the bottom step of the stairs, he knew exactly what was wrong. Ten minutes later, he came back up and handed me the bill for $120. I wasn’t paying for 10 minutes of work—I was paying for 20 years of expertise. He knew the common failure points, how different manufacturers’ machines break down, and how to diagnose the issue instantly. That knowledge—his ability to recognize patterns, categorize problems, and infer solutions quickly—is an ontology. It’s a structured way of organizing domain expertise. AI can work the same way. It doesn’t need to observe everything from scratch because someone already knows how things work. What Is an Ontology in AI? An ontology is a structured knowledge system that helps AI understand relationships between concepts. Instead of relying on massive amounts of data, AI can leverage pre-existing expertise to: Make faster, more accurate decisions Reduce reliance on large training datasets Identify patterns without requiring constant observation For example, a sales expert already knows: The 10 essential steps to close a deal The 4 types of messaging that resonate with buyers The common objections and how to counter them By structuring this knowledge into an AI ontology, teams can fast-track AI product development by enabling the system to: Make informed recommendations based on established best practices Ensure consistency and reliability instead of guessing Spot gaps in existing strategies and suggest improvements How Ontologies Improve AI Decision-Making One of the biggest risks in AI is hallucination—when AI generates inaccurate or nonsensical results. Ontologies prevent this by: Providing a structured framework that AI can rely on Ensuring AI sticks to validated knowledge instead of making random associations Allowing human experts to refine and expand AI knowledge over time This human-in-the-loop approach creates a continuous learning loop, improving both AI accuracy and business insights. Pre-Built Frameworks: The Shortcut to AI Product Success Most AI applications aren’t solving completely new problems. They’re tackling challenges that people have already structured into frameworks. If you think about cooking there are known techniques—grilling, roasting, sautéing. You don’t invent a new way to cook every time. Instead, you apply proven methods and focus on what makes your dish unique. The same principle applies to AI product development. Pre-built AI frameworks allow teams to: Reduce development time by building on existing knowledge Improve AI reliability with structured methodologies Accelerate time to market by skipping unnecessary trial and error Choosing the Right AI Framework When integrating AI ontologies into your product, consider: Alignment with your core problem – Does the framework fit your AI use case? Flexibility and customization – Can it adapt to your needs? Industry support and documentation – Is there a strong knowledge base behind it? Humans in the loop - Do you have access to experts to help AI Reliability Through Ontological Design By incorporating ontologies, AI product teams can move faster, avoid unnecessary guesswork, and build more reliable systems by: Spotting gaps before they become problems Reducing the risk of AI generating inaccurate outputs Providing teams with a structured approach to development This approach isn’t about limiting creativity—it’s about focusing innovation where it matters most. Build Smarter, Move Faster Leveraging ontologies and pre-built AI frameworks allows businesses to: Accelerate AI development without starting from scratch Ensure reliability and consistency in AI decision-making Optimize efficiency and focus on innovation With structured knowledge guiding AI, teams can get to market faster, improve product quality, and drive better business results. Image: Took this in the glass-ceilinged Robert and Arlene Kogod Courtyard at the National Portrait Gallery in Washington DC.

The first few weeks of any product development effort are tricky. You are trying to figure out what you’re doing and how to work together. The expectations are always high and the timeline short. Recently I discovered the premortem in the excellent book  Scaling Up Excellence: Getting to More Without Settling for Less  by Robert Sutton and Huggy Rao. I’m a huge fan of Sutton’s work . The premortem has been a great tool when we’re working with prospects or new clients. Prospective hindsight, imagining that an event has already occurred, increases the ability to correctly identify reasons for future outcomes by 30%. It was created by Gary Klein and popularized by Nobel winner Daniel Kahneman . This is a helpful exercise because We have an optimistic bias and tend to be really bad at predicting the future. Compare gym goers in January and March and you’ll see what I mean. It creates an opportunity for team members to feel valued and share their knowledge while preventing them from being labeled as pessimists. It predisposes the team to look for signs of trouble and take early action What I’ve found in the few times that I’ve done it is that while what success looks like is different, how we get there and the challenges are pretty similar. Clarity of purpose, clarity of priorities, shared commitment and good communication are critical. The premortem creates a moment when we all acknowledge the risks without pointing fingers or being pessimistic and we realize that we are in this together and we need to be open and honest to make it work. How I run it Break into team unmitigated disaster and team roaring success Each team comes up with a story behind using stickies. It’s important that the roaring success team is very specific about what that success looks like. In Sutton and Rao’s book they talked about a team that created a press release with metrics. Stickies should be grouped in clusters and labelled Teams report their outcomes and discuss how they can achieve roaring success and avoid the unmitigated disaster. Resources Scaling Up Excellence: Getting to More Without Settling for Less The Power of Premortems  Risk registers and meetings are boring – try a project premortem instead … Performing a Project Premortem Image: The statue of Avram Iancu  in Cluj-Napoca, Romania