HubSpot: Only 6% of firms see transformational AI value despite 90% adoption

Almost every company is using AI. Almost none are getting much out of it. That's according to Hubspot, which surveyed thousands of companies across multiple industries. In her keynote this morning at UNBOUND,, CEO Yamini Rangan said 90% of the companies Hubspot surveyed are using AI. But only 6% are seeing transformational results.

"That's such a huge gap," Rangan said.

That doesn't mean the other 94% are getting nothing. But it does mean most companies are spending on AI without seeing much change in the numbers that matter: demand, deals and customers. Even worse, AI with poor data and minimal context does more harm than no AI at all, resulting in poorer business performance.

HubSpot has a product answer to that, of course. At the same event, the company launched what it calls its most significant release in years: a CRM that updates itself, a rebuilt Breeze AI assistant that takes action instead of just answering questions, a new hub for managing the business context AI runs on, a new on-demand AI agent builder for companies and the first CRM integration with ChatGPT Ads.

All of it is built on a single premise: AI is only as good as what it knows about your business. But the big question is: what do the businesses that generate value from AI do differently?

According to Rangan, the 6% of businesses that win with AI do three things.

1. They pick outcomes, not use cases

"First, transformational companies are solving for outcomes," Rangan said.

That often means doing less. While there's always a new foundation model, new AI-powered tool, new AI-influencer process, the best performers do less, not more. HubSpot tracked 50-some AI use cases across marketing, sales and customer service. The best performers run only four or five on average.

"Transformational companies are doing fewer use cases, not more," Rangan said.

They also pick different ones. Hubspot plotted every use case by how popular it is against how much impact it drives. The things most people do with AI — producing content, sending emails, prepping for meetings — scored low on impact.

"We get this beautiful slope moving in the exact wrong direction, away from impact," Rangan said.

That makes sense. Those use case look like easy low-lying fruit: type a few prompts, get a blog post or an email. But it's exactly what anyone can do with a few prompts of their own, and the bigger problem is that most tools lack context that will make those emails or blog posts rich and impactful: customer knowledge, product knowledge, industry knowledge.