A manual message audit takes a PMM about six hours and finds maybe sixty percent of the contradictions. The AE deck still says "platform for revenue teams" while the homepage now says "AI agent for B2B sales." The pricing page calls it a "workspace." The CEO's last podcast called it an "operating system." None of these are wrong individually. Together they're four products.
The reason this happens isn't laziness. It's that surfaces drift independently — the website is owned by marketing, the deck by sales enablement, the pricing page by product, the CEO's voice by the CEO. By the time anyone notices, the contradiction has been live for two quarters.
An analyst — meaning a configured AI agent with a positioning brief, a corpus of your current surfaces, and a structured prompt sequence — does this work in roughly twenty minutes. Not because the model is smarter than your PMM. Because it doesn't get bored on page forty-seven of the sales deck.
What an analyst actually checks
The model isn't reading for vibes. It's running a structured comparison across a fixed set of axes you define up front. Get those axes right and the output is sharp. Get them lazy and you'll get a generic "your messaging seems mostly aligned" — which is what you'd write yourself if you didn't care.
The five axes that catch the most drift in our client work:
The five consistency axes
Every surface gets scored against these five. The output is a matrix: surface on one axis, the five lenses on the other, with the actual quoted language in each cell. The contradictions become visible the second you put the matrix on a screen.
How to run the check
What the output should look like
The deliverable from a single check isn't a report. It's a working document a PMM can hand to the relevant owner of each broken surface that afternoon.
The right division of labor: the analyst finds the contradictions and quotes them. The PMM decides which contradictions matter and what to do about them. Don't hand the strategic call to the model. It's bad at it, and you'll resent the output.
Where the model gets it wrong
Three failure modes worth naming up front, because the first time you see them you'll lose trust in the whole approach.
What to do Monday
Pull the last three pieces of public copy your team shipped — a homepage update, a pricing change, the most recent sales deck. Run them against the five axes by hand, just once. You'll find at least two contradictions in under an hour. That's your baseline. Now you know what an analyst-driven monthly check would have caught — and what it costs you that nobody currently does.
The first month I ran this, the analyst flagged that our pricing page still called it a "workspace" eight months after we'd moved everything else to "agent." I'd literally walked past that page every day.
Keep reading
7 Signs Your Messaging Is Drifting (And How to Catch It Early)
Messaging drifts the way codebases drift — each local change looks fine; the aggregate contradicts itself. Here are the seven patterns that appear first.
When to Refresh Your Positioning (Not Just Your Messaging)
How to tell whether the problem is positioning or execution — the four signals that mean the thesis is wrong, not the copy.
Positioning Audit: How to Score Your Own Work Objectively
Scoring your own positioning is structurally hard — you wrote it. Six disciplines that reduce the bias without outsourcing the audit, plus the rubric.
Analyst
AI strategy advice grounded in your own context — not generic playbooks.
The Analyst is a chat-based AI strategist that reads your Strategic Context, past audits, and competitive signals before answering. Ask it anything from 'why are we losing to Competitor X' to 'how should we reframe our pricing page' — and get answers that are actually about you.
- ✓Reads your own positioning data before responding
- ✓Grounded in audit findings and competitor signals
- ✓No hallucinated advice — evidence cited inline