Here is a marketing riddle: Take one of the best B2B brand builders alive, give him a big title, a near-trillion-dollar company to market, and the most famous product on earth to sell.
Why would that be a disaster waiting to happen?
Welcome to Colin Fleming’s new job at OpenAI.
Fleming is the real deal. Thirteen years at Salesforce, culminating as evp of global marketing. Two years as CMO of ServiceNow, working wonders. He knows how to take a complex B2B product, build a brand around it, and drive pipeline. He isn’t the kind of marketer you hire when you are serious about enterprise revenue, he is literally at the very top of the list. But he may be walking into the hardest marketing job in the world.
Fleming joins as CMO, Business—a new title, deliberately scoped to OpenAI’s enterprise push. His predecessor, Kate Rouch, stepped down in April after fighting late-stage breast cancer through her entire tenure. She led global campaigns, ran back-to-back Super Bowl ads, built the marketing function from the ground up, and did it all while undergoing treatment. An extraordinary act of professional commitment.
But OpenAI’s brand is in trouble right now. Start with the structural chaos. In the space of a few weeks earlier this year, the company lost its CMO, its COO transitioned to a vague “special projects” role, its head of product took medical leave, and a series of senior researchers walked out. The company is simultaneously preparing for an IPO while projecting losses of $14 billion by the end of 2026.
That is not a stable platform from which to build an enterprise brand.
Then there is the advertising decision. In January, OpenAI confirmed it would begin running ads inside ChatGPT’s free tier. The format is contextual, clearly labeled, separated from responses. Rationally it all makes sense. Enterprise and paid tiers remain ad-free. OpenAI insists the ads will not influence ChatGPT’s answers.
But we do not live in a rational world, even in the highly processed realm of AI. A Harris Poll survey taken in the days before launch found that 75% of Americans would trust AI shopping recommendations less if its results were sponsored. When you await answers from a blinking cursor, answers you don’t know the answer to, trust becomes the single most important asset OpenAI has, especially in the enterprise market.
Fleming understands this territory better than almost anyone. His entire career has been built on the specific challenge of selling to procurement committees, convincing CFOs, and building the kind of sustained brand credibility that closes enterprise deals.
Salesforce under Marc Benioff was the master class in B2B brand-building: clear positioning, massive share of voice, distinctive brand assets, relentless customer success storytelling. Fleming absorbed all of it, so much so that ServiceNow followed the same playbook, and might even have applied it better. He knows exactly what enterprise buyers need to hear, and how often they need to hear it.
But Fleming’s Salesforce instincts will now collide directly with the OpenAI reality.
Salesforce and ServiceNow sold trust as their core proposition. Their customers willingly passed on sensitive data, critical workflows, revenue operations. The brand promise was: we will never screw you.
That promise was credible because neither company faced a structural conflict between its consumer product and its enterprise product. OpenAI does. The free ChatGPT that runs ads is the same model enterprise customers are being asked to trust with their most sensitive internal data. That dissonance is not trivial.
Then there is the competitive context. Anthropic is suddenly the beast in the room. It is growing faster, smarter, cleaner, and it’s winning in the regulated industries—financial services, healthcare, legal—where enterprise deals are biggest and trust matters most.
Meanwhile Google’s Gemini is slashing prices and going after developer mindshare. Fleming is not entering a market where OpenAI can coast on brand halo. He is entering a knife fight with a smaller-than-expected pocket knife, and facing a guy with the machete and a woman with an unlimited arsenal of throwing daggers.
The dual-CMO structure is also a potential issue. OpenAI has split marketing into two roles: one for consumer, one for business. Presented as sophistication, it’s actually an admission of conflict. The consumer brand and the enterprise brand have diverged to the point where one person cannot credibly manage both. That is a strategy problem, and not one Fleming can easily solve.
None of this should be read as a prediction that Fleming will fail. He is exactly what OpenAI needs. If anyone can build a credible B2B brand on top of a messy underlying reality, it is someone with his track record.
But good hires do not always fix declining fortunes. They work within them, around them, and sometimes despite them.
The real question for OpenAI is not whether Fleming is talented enough for this job. The question is whether the company is willing to make the harder strategic decisions that Fleming will inevitably ask. About product separation, about the advertising model, about what OpenAI for Business actually stands for.
Hiring a marketer like Fleming is just the start of the problem being sorted.