HubSpot's CEO Just Quantified the Gap Between Using AI and Getting Results From It
Ninety percent of companies are using AI. Six percent are getting anything transformative out of it. That is the number HubSpot CEO Yamini Rangan opened with this week at the company's newly rebranded flagship event, and it is a more useful data point than most of what gets published about AI adoption, because it does not measure whether companies adopted AI. It measures whether adopting it did anything.
The Rebrand Behind the Number
HubSpot retired INBOUND, its conference brand for 15 years, and replaced it with UNBOUND at its September 16 to 18 event in Boston. The name change tracks a real shift in the product, not just the marketing. HubSpot is repositioning itself around AI agents fed by what it calls a Growth Context layer: a system meant to give AI tools the account history, buying signals, and relationship context that generic AI output has always lacked.
The framing matters more than the feature list. Rangan's 90 and 6 percent statistic was not offered as a caution against AI. It was offered as the justification for the rebrand: adoption has already happened, and it has not been enough. HubSpot is betting that the next phase of the AI market is not about who uses it, but about who built the infrastructure to make it produce something real.
That bet is worth taking seriously regardless of what any one vendor ships next. A conference renaming itself after 15 years under the same brand is a company staking its own credibility on a read of where the market actually is. HubSpot did not rename INBOUND because AI adoption stalled. It renamed it because adoption succeeded and still was not the thing that separated winners from everyone else.
Adoption Was Never the Hard Part
Ninety percent adoption sounds like a finish line. It is closer to a starting gate. Once a capability is available in nearly every tool a company already pays for, using it stops being a differentiator and starts being a baseline expectation, the same way having a website or a CRM did a decade earlier.
What Rangan's number actually exposes is that most companies treated AI as an add-on to an existing process rather than a reason to rebuild the process. They pointed a language model at their existing content calendar, existing email sequences, and existing sales scripts, and expected the output to carry more weight than the input that produced it. It did not, because it could not. A tool that generates text faster does not generate a perspective that was never there to begin with.
That is the actual content of the 6 percent gap. It is not a gap in tooling access. Every company in the 90 percent has access to roughly the same models. It is a gap in what got fed into those models, and in whether anyone built a system around a real point of view before turning AI loose on production.
This is a pattern with a track record outside of AI specifically. Every time a production cost drops sharply, whether that is desktop publishing in the 1990s or cheap video production in the 2010s, the same two waves follow. The first wave floods the market with more of what already existed, at lower cost. The second, smaller wave uses the same drop in cost to do something no one could previously afford to attempt, and that second wave is where the durable advantage sits. AI adoption at 90 percent is the first wave. The 6 percent getting transformative results are already in the second one.
What Separates the 6 Percent
The companies getting transformative results from AI are not the ones running the most prompts. They are the ones who did the harder work first: defining a specific audience, a specific set of beliefs worth defending in public, and a specific body of proof, then using AI to scale the publishing of that perspective rather than to invent one from nothing.
That is the logic Kyroiq was built on, tested on its own record before this statistic existed to describe it. Two independent ventures, one in finance and one in travel, started from zero public audience and reached more than 4,000 followers combined across platforms in two months. The pipeline behind that growth was AI-powered throughout, but the output was never generic, because the system was built around a specific point of view before a single post went out. The partner opportunities that followed arrived through inbound interest, not outreach, which is the same outcome HubSpot is now describing at a market-wide scale.
The mechanism generalizes past any one company's build. AI can produce volume at a cost and speed no manual process can match. It cannot produce a perspective. That still has to come from somewhere, and the companies skipping that step are the ones showing up inside Rangan's 90 percent with nothing to show for it.
The Practical Read for Anyone Building With AI Right Now
The lesson from HubSpot's own numbers is not to use AI less. It is to stop treating AI adoption as the achievement and start treating it as infrastructure for something that has to exist independently of the tool: a real, defensible point of view, tied to specific proof, published consistently enough to compound.
Companies still asking which AI tool to adopt are asking the wrong question. The 90 percent already answered that one, and it did not move the number that matters. The better question is what system the AI is running inside of, and whether that system was built around a perspective before it was built around output.
Ninety percent of the market just adopted the same capability. The 6 percent who are winning with it built something the other 94 percent skipped.