There’s a question going around product marketing that most of us would rather not say out loud. If AI can produce the deck, the one-pager and the battle card, does a company still need a PMM?

In the Product Marketing Alliance’s State of Product Marketing 2026 report, Mike McGee of Vantaca raises exactly that question. It’s a fair one. And the report’s own data points to a better answer than “yes, because we’re creative.”

The work was never the bottleneck

Look at how product marketing teams are actually built:

• 29% of PMMs work alone. About 8 in 10 are on teams of five or fewer.

• 35.1% support six or more products. 44.6% support three or more product managers.

• 63.2% still own competitive intelligence as part of the job.

That isn’t a team short on content. It’s a team short on time. Most PMMs could write another battle card this afternoon. What they can’t do is be in every deal, answer every rep’s question as it comes up, and keep every asset current across six products.

So when AI makes content faster to produce, it doesn’t remove the PMM’s job. It removes the part that was never the constraint.

The real risk is becoming a production house

Several contributors to the report describe the same trap. PMM teams get pulled into whatever needs a deck this week, and turn into an internal production house: a help desk for content requests.

AI makes that trap worse, not better. If the PMM’s value is measured in assets shipped, then a tool that ships assets faster really does compete with the PMM. The PMMs who should worry are the ones whose job has quietly become production.

Where the value is moving

The report shows what companies now expect product marketing to deliver:

• Win rate as a PMM KPI rose from 28.1% in 2024 to 37%.

• Sales confidence rose from 18.8% to 30.3%.

Those aren’t content metrics. They’re about what happens in the deal: whether the rep says the right thing, with confidence, at the moment it matters. That’s judgment, not production. Knowing which proof point beats which objection, which claim is safe to make, and which competitor move actually matters this quarter.

AI can generate words. It can’t decide what your company should say. That still takes someone who knows the market, the product and the buyer.

AI needs a source of truth

The report also shows why AI alone isn’t the answer. 80% of PMMs use AI regularly, but concern about accuracy and hallucinations is the top barrier to using it, cited by 63.6%. And only 6.9% say formal processes or guidelines best describe how they use it.

If that’s how PMMs use AI, imagine how reps use it. They’re asking general-purpose tools about your product and your competitors, and getting answers nobody approved.

The fix isn’t less AI. It’s AI grounded in content the PMM owns. When the PMM’s judgment becomes the source every answer draws from, AI stops competing with product marketing and starts carrying it into deals the PMM could never attend in person.

So what is the PMM for?

The PMM is the source of truth. Not the person who makes the battle card, but the person who decides what’s true in it, and makes sure every rep gets that version.

The PMMs who come out ahead won’t be the ones who produce the most. They’ll be the ones whose judgment shows up in every deal, whether or not they’re in the room.

That’s the idea behind Marnie: reps ask in Slack and get answers from the content their PMM has approved, so the PMM is present in every deal without being pulled into every thread.

Figures from the Product Marketing Alliance, State of Product Marketing 2026.