What New HBR Research Says About AI and Revenue: Our Take

AI has made revenue teams faster. It hasn’t made them much money yet.

That’s the clearest finding in a new pulse survey from Harvard Business Review Analytic Services, sponsored by Seismic. In June 2026 HBR surveyed 208 people at B2B companies who know how their revenue teams work. 75% say their company has increased its use of AI in sales, marketing, support or enablement over the past year.

The returns are uneven:

• 52% say AI has made their processes more efficient.

• 11% say it has grown revenue.

• 8% say it has shortened the sales cycle.

• 7% say it has raised win rates.

Time saved is real. Revenue gained is rare. The report’s explanation is the part worth reading. The gap isn’t about models or budgets. It’s about the basics. Five stand out.

Know Your Business Before You Automate It

The top barrier to getting more from AI isn’t technical. It’s strategy. 45% of AI adopters say their company lacks a clear AI strategy. Data problems (30%) and budget (24%) rank well below it.

Steve Busby, CEO of Revenue Operations Associates, told HBR that leaders should picture the business three years out. Then they should plan the operating model, the org chart and the go-to-market plan that get them there.

That starts with questions that predate AI. Who do you serve? Where are you trying to go? What do your best customers need from you? Teams that can’t answer these clearly end up automating whatever is easiest. The report finds that too: 32% say AI is mostly used for simple tasks rather than high-impact work.

Give AI the Context It Doesn’t Have

A general model knows a lot about the world. It knows nothing about your accounts, your pricing, your last lost deal or why your best customer renewed.

The survey shows where that leads. Real-time AI is mostly used to make things: meeting prep (54%) and call summaries (50%). Few teams use it to guide what happens next. 19% use it to recommend next-best actions. 18% have AI assisting reps in real time.

Recommending the right next step takes context. Who the buyer is, what they did last week, what worked with similar accounts. Without that, AI writes faster but doesn’t decide better.

Judgment Is the Scarce Skill

Asked which skill matters most for getting value from AI, 55% of adopters picked critical thinking and judgment. That ranked ahead of knowing when to use AI (49%) and AI literacy (35%).

Confidence lags. Only 12% say their revenue teams are highly confident using AI well. 55% report low confidence.

AI can produce leads, drafts and recommendations quickly. As the report puts it, it can’t reliably tell which new use cases matter most or when customer preferences are about to shift. That’s still the job of people who know the customer.

Decide Who Governs the AI

This is the widest gap in the report. 88% of adopters agree revenue teams need a strong governance framework to use AI safely and effectively. Only 32% say their company is building one in the next year.

Ownership is unclear too. When asked who is mainly responsible for AI in revenue work, adopters most often named the CIO, CTO or chief data officer (21%), then the CEO (15%). The CRO came in at 12%. Sales leaders and enablement leads barely registered.

Regal Rexnord offers one model. Its AI governance committee meets every two weeks and pulls in security, legal, privacy, compliance and the business. Each use case gets reviewed before it goes live. As AI starts acting on deals, not just drafting emails, someone has to decide what it may do and check that it did it right.

Measure Revenue, Not Hours Saved

36% of adopters say they struggle to measure AI’s value. Elodie Blyth de Fontenay, head of data and AI transformation at Royal London Asset Management, told HBR the return “can’t be tied directly to just an increase in profitability.”

Hours saved are easy to count, so that’s what gets reported. But hours saved only matter if they go somewhere. The better questions: did reps spend the time on the right accounts? Did win rates move? Did deals close sooner? Gabe Sweet of ZoomInfo summed up the shift: the question is no longer efficiency. “Now it’s about, ‘How do we actually drive outcomes for the business?’”

Back to the Future, Back to Basics

The interfaces of revenue work are changing fast. More work is going headless. Agents do it in the background, and people review the results instead of clicking through screens. The tools you use today won’t look the same in two years.

What holds is knowing your business. Who you serve. Where you want to go. What your customers need. Who decides what the AI does, and how you’ll know it worked.

That’s the idea behind Actively. We call it GPS for your GTM. A GPS is only useful if it knows two things: where you are and where you’re going. Generic AI knows neither. It has read the internet, not your pipeline.

So we’re building Actively to learn how your business works. Your accounts, your wins and losses, the way your team sells. The goal is AI that points reps to the right next step, for your business, with people setting the direction and checking the route.

Source: Harvard Business Review Analytic Services, “Unlocking AI’s Revenue Potential,” a pulse survey sponsored by Seismic. 208 respondents at B2B organizations, fielded June 2026.