It's not better prompts or fine-tuned models. It's something that exists inside your organization right now that your competitors don't have access to and never will.

I'm calling it the canonical layer.

When you generate a competitive positioning brief with AI, it pulls from public information. Analyst reports, review sites, conference talks, blog posts, LinkedIn thought leadership. The accumulated category language that's been published over the last decade. That corpus is shared, which is why the outputs converge across every vendor in your space.

There's another corpus AI doesn't have. The proprietary context that only exists inside your organization. The things that never get published, never make it into a G2 review, never show up in a Gartner report. Real customer language.

Click to zoom

The actual words buyers use in discovery calls when they're trying to explain the problem they're solving. The metaphors that work. The objections that come up in every deal and the answers that actually move them forward. Win/loss patterns your competitors don't see.

Why you win against Competitor A and struggle against Competitor B. What buyers say when they choose you and what they say when they don't. The gap between what your product team thinks is differentiated and what actually closes deals.

Product decisions and their downstream implications. Why you architected something a certain way and what that unlocks six months from now. The technical tradeoffs that don't show up in a feature comparison chart but matter in implementation. Sales conversations that never make it into the CRM. The objections reps hear and don't log. The questions that come up in every demo.

The moments where a deal stalls and nobody's quite sure why. Field feedback that lives in Slack and dies there.

What customer success is hearing in renewals conversations. What solutions engineers are building workarounds for. What's breaking in ways product hasn't prioritized.

This is the context that makes your positioning different from everyone else's because you have information your competitors don't.

When everyone has access to the same production capabilities, the only competitive advantage left is what you feed those capabilities. If you're generating positioning with AI and your competitor is generating positioning with AI, and you're both pulling from the same public corpus, your outputs will converge.

Click to zoom

If you're feeding your AI proprietary context, your output diverges because the input is different. That's what differentiation looks like now, and it's in what makes the artifacts possible.

Most PMMs aren't set up to build a canonical layer because we're set up to produce content. Our calendars are full of launch decks and sales enablement and campaign briefs, and the work that builds visibility is artifact production, so that's where the time goes.

Building a canonical layer requires different work. Less time in slide decks and more time in sales calls, win/loss debriefs, customer conversations, Slack threads where the field is working through objections in real time. It's extraction work, and it doesn't create an artifact you can point to at the end of the day.

That's a problem when your performance is still being measured by output velocity. The output layer is commoditizing, which means the artifacts aren't the strategy anymore. The canonical layer is. The question is whether you're building it intentionally or just hoping the context shows up when you need it.

Next week I'll talk about what building a canonical layer actually looks like in practice. As a system.

I'm speaking at the Product Marketing Alliance Seattle Summit on June 17th about what PMM is actually for when the output layer commoditizes. We'll talk about building canonical layers in practice, not just theory. If you're going, come find me and lets chat.