Where Do You Sit?

There's a product marketer — let's call her Sarah — who runs competitive intelligence for a mid-market cybersecurity company, and if you watched her work for a full day you'd swear she was one of the most sophisticated AI users you'd ever seen. She has Claude open in one tab and Perplexity in another and a custom GPT she built specifically for parsing earnings call transcripts, and by four o'clock she's produced a competitive battlecard that would have taken her predecessor a week and a half. She's fast. She's thorough. She's convinced she's at the bleeding edge of AI-augmented product marketing. And she might be wrong about that last part — not because she's using AI badly, but because she's never once let it inside the room where her actual thinking happens.

Then there's Marcus, a PMM lead at a Series C infrastructure company who started using Claude for positioning work about eight months ago and somewhere along the way crossed a line he didn't notice at the time. He doesn't draft positioning and then ask AI to sharpen it. He thinks through positioning in conversation with AI — feeding it raw customer interview transcripts, exploring framings he hadn't considered, letting the model's responses redirect his strategic reasoning in real time. His competitive narratives are better than they've ever been, his launch plans have a creative ambition they never had before, and when you ask him to describe what he does for a living, he can't quite separate the PMM part from the AI collaboration part anymore. His professional identity has — quietly, without drama — reorganized around the relationship.

Sarah and Marcus are both power users. Both would check the "advanced AI adoption" box on any survey. And in the flatness of most industry research, they'd be treated as essentially the same data point. But they are living in fundamentally different relationships with artificial intelligence — and the difference matters more than almost anything else in the current conversation about AI and the future of work.

The Three Dimensions

Edwards proposes three dimensions. Each one is a yes-or-no question about your relationship with AI, and each one is doing different work.

The first is Cognitive Permeability — whether AI is genuinely inside your reasoning process. Not whether you use it. Whether you think with it. CP=1 means iterative back-and-forth dialogue where AI's responses redirect your thinking in ways you didn't plan. CP=0 means you query, you receive, you review, you move on. Your thinking happens before and after the AI interaction, not during it. For a PMM, this is the difference between drafting a positioning statement and then asking Claude to "make it punchier" versus working through the positioning logic in dialogue — exploring competitive angles, pressure-testing value prop hierarchies, discovering framings you hadn't considered — and letting those discoveries actually change where you land.

The second is Identity Coupling — whether your professional self has reorganized around AI. This is the dimension that surprised the researchers most, and the one that matters most for the conversation we're about to have. IC=1 means you genuinely cannot describe your professional identity without reference to AI. Your workflow, your sense of competence, your career trajectory have all restructured around the relationship. IC=0 means your professional identity remains conceptually separate from AI, even if you use it constantly and even if you call it transformative. Only 11.8% of Edwards' sample crossed this threshold. The headlines about AI fundamentally reshaping professional identity may be — for now — describing a much smaller population than anyone assumed.

The third is Symbolic Plasticity — whether the meanings and success criteria you apply to your work have become fluid. SP=1 means you're renegotiating what "good" looks like — exploring multiple framings, redefining quality standards, discovering that the form factor of a deliverable or the metrics of a launch might need to be something fundamentally different from what you've always assumed. SP=0 means your standards are anchored. Good positioning is good positioning. A successful launch looks like what a successful launch has always looked like. The tools changed, the benchmarks didn't.

Three binary dimensions, eight possible combinations, eight distinct profiles. And the distribution isn't what you'd expect.

The Two Dominant Patterns

Nearly two-thirds of the professionals in Edwards' dataset occupy just two roles: the Framer and the Doer. Framers — 33.4% of the sample — let AI into their reasoning while keeping their identity and standards stable. Doers — 30.3% — keep AI entirely outside their reasoning process, using it as an execution layer while all the strategic thinking happens in their own heads. Together, these two patterns account for the vast majority of sophisticated AI users. The identity-coupled roles — the ones where your professional self has genuinely reorganized — account for just 11.9% combined.

For the PMM function, this distribution tells you something important about where the profession actually sits right now, as opposed to where the conference keynotes say it sits. Most product marketers who are serious about AI have either let it into their thinking while staying fundamentally the same PMM they were three years ago, or they're using it extensively as a production accelerator while keeping their strategic reasoning walled off. Both of these are defensible postures. Both preserve what Edwards calls cognitive sovereignty — the capacity to author your own thinking in the context of AI — at reasonably high levels.

But the really interesting story is at the edges.

The Sovereignty Paradox

The comparison that anchors the entire research is between two roles that share the same identity profile but diverge on one critical dimension. Co-authors and Creators both have IC=1 and SP=1 — their professional identities have reorganized around AI, and the meanings of their work are actively in flux. The only difference is Cognitive Permeability: Co-authors have opened their reasoning process to AI. Creators haven't.

That single dimension — whether AI is inside the thinking or outside it — creates a 1.07-point gap on cognitive sovereignty. Co-authors score 8.54 out of 9, the highest of any role in the dataset. Creators score 7.47. Same identity disruption, same meaning fluidity, radically different experiences of what all that disruption feels like.

The Co-authors in Edwards' data describe something that any PMM who's crossed this threshold would recognize immediately — the experience of iterative strategic collaboration where you're not using AI as a tool but thinking alongside it as a genuine cognitive partner. They customize AI to their needs. They maintain editorial control and creative ownership. They describe themselves using words like director and creative partner. And crucially, they use language of empowerment: the integration feels like gaining a capability, not losing one.

The Creators tell a different story. They're concentrated in roles where AI directly displaces their core output — content creators, writers, editors — and the language in their transcripts is qualitatively different. Where Co-authors say extension of myself, Creators say my job will be almost completely replaced. Where Co-authors describe partnership, Creators describe threat. Their professional identities reorganized because AI forced the reorganization, not because they chose to build a new relationship with it.

What This Means for PMMs

If you're a PMM and you're reading this, the question you should be asking yourself isn't "am I using AI enough?" — it's "which side of this divide am I on?" Because the divide isn't about adoption levels. It's about whether the most important thing AI is doing in your professional life is entering your reasoning or just entering your workflow.

This connects directly to the 10x PMM thesis we've been developing throughout the Future of PMM curriculum: the sovereignty paradox is the empirical backbone for everything we've been arguing prescriptively.

A 10x PMM is not someone who uses AI ten times as much. A 10x PMM is someone who has let AI into their reasoning process, maintained authorship over the strategic judgment that makes the work valuable, and — in doing so — expanded what one person can accomplish to the point where the old headcount math doesn't apply anymore. That's the Framer at minimum and the Co-author at the apex: the product marketer whose thinking, identity, and standards have all engaged with AI, and who came out the other side more sovereign, not less.

Two Patterns Worth Watching

Edwards found something else that deserves attention in any conversation about PMM teams and organizational AI adoption. The exceptions to the unified sovereignty pattern map perfectly onto archetypes that any PMM leader would recognize on their own team.

The first fracture pattern is seeing without acting — high awareness, collapsed agency. This is the PMM who reads every AI trends newsletter, can articulate exactly how the profession is shifting, understands intellectually that agentic workflows will restructure the function — and hasn't changed a single thing about how they actually work. They can read the story of what's happening to them. They can't pick up the pen.

The second fracture pattern is owning without seeing — high accountability, lower awareness. This is the conscientious PMM who's adopted a handful of AI tools, takes full responsibility for every piece of output, quality-checks everything, would never blame AI for a bad deliverable — and has never paused to examine how those tools might be reshaping their strategic instincts, their framing habits, their default assumptions about what a competitive narrative should even look like. They hold the pen. They're not watching what the pen is doing to their handwriting.

If you lead a PMM team, these two patterns are worth diagnosing — not with a formal instrument, but with a conversation. The seeing-without-acting PMM needs a forcing function: a project that requires them to use AI as a reasoning partner, not just an awareness source. The owning-without-seeing PMM needs a mirror: someone to help them notice the ways their cognitive patterns have already shifted, so they can make those shifts deliberate rather than invisible.

A Note on Limitations

I should be honest about where I'd push back on Edwards' framework, because intellectual honesty is the whole point of what we're building at Future of PMM.

The paper treats Symbolic Plasticity as a binary — your professional meanings are either in flux or they're not. For product marketing, I think that's too blunt. A PMM whose definition of "good positioning" is in constant existential flux is probably just disoriented. But a PMM whose understanding of how positioning gets built is evolving — who's discovering that competitive analysis can be more iterative, that customer segmentation can be more dynamic, that narrative testing can happen in hours instead of weeks — that's productive plasticity. The standards haven't moved. The process for meeting them has been blown wide open.

The other honest caveat is that the sovereignty measure compresses at the top. Nearly half of Edwards' sample scored 9/9, which means the instrument can't differentiate among the large majority of people who are doing reasonably well. The self-assessment below inherits that limitation. It'll tell you which of the eight roles you currently occupy, and that's useful. But it won't give you a precision score on sovereignty. Consider it a conversation starter, not a diagnosis.

Take the assessment. Find out where you sit. Then ask yourself whether you're there by choice.