Episode 271: The Gap Between AI Adoption and AI Strategy

Creating great products isn't just about features or roadmaps. It's about how organizations think, decide and operate around products. Product Thinking explores the systems, leadership and culture behind successful product organizations. We're bringing together insights from multiple product leaders pulled from past conversations to explore one shared topic offering different perspectives and lessons from real world experience. I'm Melissa Perry and you're listening to the Product Thinking Podcast Podcast by Product Institute. Hello and welcome to another episode of the Product Thinking Podcast. Today, this is a different kind of episode. There's no guest today. It's just me sitting in front of a microphone telling you what we learned from the survey we just published. And then I have a little bit of news about what's next for the show. The survey is called State of AI in Product 2026 and we surveyed 309 product leaders from late April through May 202026 and I co published it with Product Circle, the practitioner community Dra Moita and Sergey Lazar have been building over the last couple of years. The full report is out and the link is in the show notes if you want to read it. But I'm going to tell you a little bit about why we ran the survey and what we found in there. We ran this survey because the conversation about AI and product for the past two years has been a conversation about tools, which assistant for which task, which model, which plugin. That conversation isn't wrong, it just isn't the most useful one anymore. Almost everyone has the tools in some shape or form. The question I wanted to answer is whether the tools are changing, how the work actually happens. So we asked 309 senior product people across 40 countries and concentrated in the US and Europe about what AI is doing inside their organizations right now. These are the five major findings. Let me walk you through them. The first finding is that the tools have changed, but the work did not. The first thing we learned is the spine of the whole report. 87.7% of respondents say that their organization uses AI coding assistance in product work. 85.4% use AI tools for product work like research, writing and analysis. 71.5% have built internal AI tools, 69.9% have shipped AI powered features in their product. So what we see here is that adoption is happening. It's everywhere. But then we asked whether AI is actually strengthening their product operating model and only 36% of people said yes. About a quarter of people said that it's exposing weaknesses that were already there and 6% said it's actually making things worse. This is what I suspected from my own conversations, and I was curious to see how it scaled. So what I'm taking away from this is that the tools are almost everywhere. The operating model, around the tools, the way decisions get made and reviews happen, and customer signal travels that hasn't caught up. That's the gap the rest of the findings are really looking at too. So let's go to finding number two. The bottleneck is still upstream. The second finding is not new for me, and it's the most operationally useful one for me. Though. About half of the respondents say AI is having a high impact in engineering and development. 45% say the same about design and prototyping. But then it drops fast. Strategic planning, qa, customer research, cross team collaboration, all of those were in the single or low double digits. 1pm in the survey wrote a line I'm now quoting to every leader I've spoken to since. He said delivery of designs and code got very fast. Delivery of good decisions became the new bottleneck. Now, I don't think this bottleneck is new. It's exactly what I talk about in the build trap. What I've been seeing is that AI has now just made that build trap bigger. We can deliver more code faster, we've got more prototypes that can be made faster. But what are we building? Is it being used? Those upstream issues have always been the issues and AI is not making them better. For now, hopefully in the future we have things that can actually help with it. But for now, sometimes it's actually making it worse. So if you're a product leader looking to where should you invest in your team next? It might not be in another coding assistant or prototyping tool. Look towards how you can use AI in the discovery work. The prioritization, the review cadences, that's the work that happens upstream of the build. There are a lot of AI tools on the market now that helps with data analysis, customer feedback aggregation. I would really look to see how you can incorporate those in discovery processes or revisit your entire discovery process, the way that you make decisions as a whole, and then go figure out how can you make that faster. Our third finding is that AI is a multiplier, not an equalizer. The third finding is the most uncomfortable one. For organizations that hoped AI would be the great equalizer, and it isn't. In the sample, it looks more like a multiplier. Organizations that rated their operating model as mature and healthy before AI are 1.7 times more likely to say AI is strengthening it and they're 3 times less likely to say AI is making things worse. That's pretty good. And the size cut is the cleanest signal in the data set. Teams of 1 to 50 people report that AI is strengthening their operating model at 48%. Organizations of 500 or above collapse down to 20%. Even though the larger organizations are investing more, they're running more formal training, they're hiring more for AI, they're standing up more programs, they're spending more, they're just converting less. This is really telling me what I suspected about operating models here, too. Whatever the operating model was before AI, AI just made more of it, and that's the multiplier function of it. So if you didn't have a complete operating model, if it was still broken, it's just going to make it worse. So now that brings us over to finding number four, which I think goes hand in hand with this. That's the strategy gap. The fourth finding is the biggest single axis gap in the entire data set, bigger than any difference by tool, industry geography or company size. And it's not about technology at all. It's about strategy. Specifically, it's about who thinks one exists. 62% of the product managers in the survey cite that there are no clear AI strategies at the top as a big challenge for their company. Now, among the C level respondents, that number drops to 19%. So that's a 43% gap there. This matters. It probably isn't that executives have no AI strategy. It's that whatever strategy exists at the investment level, the level that the CEO and the board are really thinking about, this hasn't been translated into the operating model rules that a PM needs to do their job on a Monday morning. It doesn't go into their priorities, it doesn't go into their decision rights. It doesn't go into the review virtuals, what to use AI actually for, what not to use AI for, who reviews the output. The strategy is up at the top. It exists, but the translation is missing. How do I take this strategy and actually apply it to my own work? That's what's being left out. So that's a big finding for finding four. If you. And here I would really encourage you to start looking at how can I make my AI strategy more explicit? How do I make sure that people understand it and how to actually apply it? That's what makes strategy actionable and that makes sure that people can actually take advantage of it. Our last finding, finding number five, is the fear of moving too slow. So this is what product leaders are afraid of right now. When we surveyed them, we found that the biggest fear FL for 35% of respondents was moving too slow and falling behind. Only 19% said that their biggest fear was the opposite moving too fast without enough guardrails. So the dominant fear in product leadership today is being late, not being reckless. Over 28% of respondents, though, pick something completely different. Their fear was over investing in tools without changing how the organization works. And this one is the warning I'm taking away from the report. The fear of being slow is shaping the next round of tool purchases and the second order risk of those purchases, money spent without redesigning the entire system, is already visible in the data. So what does this mean for you if you are leading a product organization right now? Here's a few things that I think you can look at the next AI investment in your org is not another tool. It's a workflow redesign. Audit how decisions move. Audit what your teams are reviewing. Audit how customer signal travels through the organization and make that operating model legible before AI surfaces it on your behalf. Translate that AI strategy that exists at your level into operating roles your product managers can actually use. If your product managers can't tell you how they should be incorporating AI into their work, concretely, this translation layer is not happening well. Stop measuring AI adoption as your main metric. Measure cycle time. Measure decision quality. Measure customer insight velocity. Measure outcomes. Adoption is the input. This is not the goal. Right? Everybody's trying to adopt AI in order to reach those outcomes. So let's start trying to figure out how to measure those outcomes well and at the end of the day, train product design and engineering together. Function specific tool training widens the fluency gap. Cross functional training really compresses it. Now if you are a product manager listening to this, the data is less comforting in one way, but a little bit more useful in another. The less comforting part is that most of you are working inside an organization where the AI strategy exists somewhere above you in a forum that hasn't really reached your work. Now that's not your fault, but it is an opportunity. And that's where the more useful part of this is. The product managers who will pull ahead in this error are the ones who can connect strategy to daily decisions. Now that's not a tool skill, it's a systems thinking skill. Decision quality customer insight. The ability to walk into a strategy meeting, tell senior leaders what their AI bets mean for what the team should and shouldn't ship next quarter. All those things really matter. So this is a skill that compounds and this is really where you should invest when you're figuring out what should I learn, what should I get better at? So that's a little bit of a wrap up about our survey. And before I close out this, I want to tell you a little bit of piece of news about the show. I'm going to take a break from new podcast episodes for a little while. There's a lot changing here right now. The next cohort of CPO accelerator is going to kick off in September. We've got some new things happening at Product Institute. I've got this restaurant that I built and I've got a couple other projects that I'm not ready to talk about yet. But it's a good kind of busy and it's also the kind where I'd rather pause the show than do it halfway. So this is the last new episode for now. We'll be back. I don't have a specific date yet, but I'm not going far. You'll still find me on LinkedIn. And if you want to read the full report on the state of AI in product 2026, it's at productcircle co state of AI2026. The link is in the show notes@productthinkingpodcast.com and in the meantime, the back catalog of our podcasts are still there for you. If you have never listened to the conversation with Matthew Skelton or the episode with Denise Tills on Product Operations, those will be good ones to revisit while we're on pause. So thank you so much for listening. Thank you for the questions you've sent me over the last four years. We'll talk again soon. This is not the last time you'll hear from me, but until next time, I hope you keep escaping the build trap.