In a world where everyone claims AI, differentiation is the only question that matters.

Product marketing at IBM has always been a strategic function because the portfolio is large, the buyers are sophisticated, and the sales cycle is long. In the AI era the work has expanded. We're no longer marketing a feature inside a category buyers already understand. We're marketing a new category of work (enterprise AI, AI agents, model orchestration) into organisations that are figuring out the category at the same time we are. That changes the brief.

The product marketer in this era is part storyteller, part economist, part change-management consultant. The storyteller piece is the obvious one: we have to make the technology legible. The economist piece is harder: we have to help the buyer build the business case, because the procurement function won't approve the spend on a story alone. The change-management piece is the one most product marketers don't think of as their job: enterprise AI fails when adoption fails, and adoption fails when the organisation isn't ready. Helping the buyer prepare the organisation is part of the work now.

A lot of the headlines about enterprise AI ROI is disappointing are not really about the technology. They're about chaos. Poor data foundations. Fragmented processes. Tools that don't integrate. Teams that aren't ready to change how they work. The technology is doing what it's supposed to. The organisation around it is the constraint.

The implication: the conversation about AI ROI has to start with the data foundation and the process redesign, rather than starting with the model. We've found that when we lead with the data and the process and use the model as the proof point at the end, the business case lands and the deployment sticks. When we lead with the model, customers get excited, run a pilot, hit the chaos layer, and the pilot doesn't scale.

A Ferrari customer story we use internally: we worked with Ferrari to build the fan experience around the brand and the racing season. The customer didn't buy an AI agent for content personalisation. They bought the transformation of how their fans engage with the team across the calendar year. The technology underneath was AI, data integration, content orchestration. The story we told was the transformation. The technology supported it.

The lesson generalises. Buyers in the AI category don't want to be sold a capability. They want to be sold a future. The product marketer's job is to make that future legible, then back it with the technical proof points the buying committee needs.

Outcome first. Transformation second. Capabilities third. Most product marketing decks (especially in enterprise B2B) invert this: capabilities first, then maybe a slide on transformation, then maybe an outcome metric buried near the end. The reorder is uncomfortable because product teams want their capabilities front and centre, but the buyer is paying for the outcome. The discipline is to lead with the outcome and let the capabilities support the case.

Enterprise procurement is unforgiving. A great story gets you in the room. A solid business case keeps you in the running. A defensible TCO and ROI model gets you to signed. Product marketing has to build all three. Increasingly we're building procurement-grade business case templates that the buying customer can take to their own CFO and make our case for us. The product marketer who hands a customer a story without the financial case is leaving the deal half-finished.

The data has to be honest. Customers (and analysts) see through inflated ROI numbers immediately. We work with customer success and analyst-relations teams to build defensible numbers grounded in real deployments, then publish them. The credibility compounds. Inflated numbers might win one deal; they lose the next ten.

AI is reshaping how product marketing teams operate, rather than only what they market. The obvious wins: content variation, message testing, briefing efficiency, summarisation of analyst calls and customer interviews. The deeper wins: faster competitive intelligence (a tool that can summarise a competitor's earnings call, product announcements, and analyst coverage in minutes used to take a team a day), and richer customer pattern recognition (analysing thousands of customer interactions to surface common pain points).

The skill that still matters most in product marketing is judgement. AI gives us draft messaging, draft positioning, draft variants. The judgement of what to ship is human. The discipline I push on the team is: use AI for the drafting and the analysis, but the final cut (the line that goes in the keynote, the angle that goes on the website, the metric that goes in the customer story) is something a human owns. The judgement is the moat.

I hire for curiosity and the willingness to be honest about what isn't working. The AI category moves too fast for anyone to know everything. The product marketer who is curious, willing to learn in public, and willing to admit when a message isn't landing is the one who builds the team and the function. The product marketer who pretends to have it figured out is the one who doesn't.

Working inside IBM as a product marketer in the AI era is a particular kind of opportunity. The brand has a 100-year history of enterprise technology. The customer base is the largest organisations in the world. The expectation is that we lead, not follow, on category-defining moves. That's exhilarating and demanding. The work is to honour the legacy while pushing the new category forward, which means the marketing has to feel as ambitious as the technology.

Build the financial muscle early. Many product marketers come from a creative or communications background and treat the business case as someone else's problem. In B2B enterprise it's never someone else's problem. Learn how the CFO thinks. Learn how the buying committee builds the case. Learn what procurement looks for. The product marketer who speaks the financial language is the one who closes the deals and shapes the category.