Your AI Strategy Has a Blind Spot: Internal Adoption

Most product leaders I talk to are obsessed with how to build AI into their products. What features to ship. What models to use. How to differentiate. Those are the right questions — but they're incomplete.

The question that gets far less attention: how is your organization actually using AI tools internally to drive efficiency?

I've spent the last few years working with other leaders at Ontra in an effort to get AI tools into the hands of as many people as possible — Cursor, Claude Code, ChatGPT, Gemini, Notion AI, and more. The results have been significant. Teams are moving faster, communicating more clearly, and finding creative solutions to problems that used to eat up hours of manual work. But the path to getting there wasn't a top-down mandate. It was something much messier, and honestly, much more effective.

Start with Curiosity, Not a Rollout Plan

When we first started experimenting with AI tools internally, we resisted the urge to create a formal program. No mandatory trainings. No company-wide "AI transformation initiative." Instead, we did something simpler: we let curious people be curious.

We identified the people who were already tinkering — the ones who were using ChatGPT on their own time, or asking questions about how Cursor worked. We gave them access, gave them room to experiment, and let them run. The bet was that if the tools were genuinely useful, those early adopters would become evangelists. And that's exactly what happened.

There's something fundamentally different about hearing "this tool saved me three hours on that report" from a peer versus hearing "we're rolling out an AI initiative" from leadership. The first one creates pull. The second one creates eye rolls. We leaned into the pull.

Build the Sharing Muscle

Early adoption only gets you so far if it stays siloed. The next step — and in some ways the more important one — was creating the infrastructure for people to share what they were learning.

We set up a dedicated Slack channel where anyone could post how they were using AI in their day-to-day work. No format requirements. No pressure to be polished. Just "here's what I tried, here's what worked and here's what didn't." Some of the best use cases we've discovered came from people in non-technical roles who found creative applications that none of us on the product or engineering side had considered.

We also ran an AI competition recently, where teams submitted their best use cases so we could learn from each other. The competition format did something important: it gave people a reason to articulate why their approach worked, not just what they built. That articulation is where the real knowledge transfer happens. When someone can explain the problem they were solving, the tool they chose, and why the output was valuable, everyone in the room walks away with a framework they can apply to their own work.

Prototyping as Alignment, Not Production

One of the more surprising wins has been in how we've used AI prototyping tools like Lovable. But probably not in the way you'd expect.

We're not using these tools to generate production-ready code. We're using them to align teams around a shared vision. When teams can spin up an interactive prototype, the conversation shifts. Instead of debating wireframes or interpreting written specs, people can react to something tangible. "Do you mean like this?" becomes a much faster path to alignment and a lot of fun as teams work together to create shared vision.

The value isn't in the code these tools produce — it's in the shared understanding they create. When everyone on a team can see and interact with a rough version of what they're building, misalignment surfaces early.

What I'd Tell Other Leaders

If you're thinking about how to bring AI tools into your organization, here's what I've learned:

Lead with curiosity, not mandates. Find your early adopters. Give them space. Let them discover what works organically. A bottom-up movement is harder to ignore than a top-down directive.

Invest in sharing mechanisms. A Slack channel, a monthly show-and-tell, a competition — the format matters less than the habit. The goal is making it easy and rewarding for people to share what they're learning.

Reframe prototyping tools. If you're evaluating tools like Lovable, Bolt, or v0, don't judge them by whether they produce production-ready code. Judge them by whether they help your teams get aligned faster. That's a different kind of ROI, and it's real.

Meet people where they are. Not everyone on your team is going to be a power user on day one. That's fine. The evangelist model works because it creates a gradient — the curious pull in the interested, who pull in the skeptical.

The Bigger Picture

There's a temptation in product leadership to treat internal AI adoption as a separate concern from your product strategy. But I'd argue they're deeply connected. Teams that are fluent in AI tools build better AI products. They have sharper intuition for what works, what's possible, and where the rough edges are. They've experienced the "aha" moments themselves, which makes them better at creating those moments for customers.

Your AI strategy shouldn't just be about what you ship externally. It should include how your teams work internally. The organizations that figure this out will have a compounding advantage — not just in efficiency, but in the quality of their product thinking.

Start with the curious people. Build from there.