I have been asking a version of this question for months, and it usually gets a laugh before it gets a serious answer. If an AI agent sources the account, books the meeting, qualifies the lead, and drafts the follow-up that closes the deal, does it earn commission?
Of course not. But that question reveals exactly where revenue execution is heading and the gap it will create. Sellers and organizations who adopt agentic AI will supercharge their potential, outperforming those who don’t. More capacity, more coverage, more closed. The question isn’t whether AI belongs in your revenue motion. It’s whether you get there before your competitors do.
In the rapid-paced AI world, agents have moved past the pilot stage. They are running in production across sales teams right now. Juniper Research counts 3.3 billion customer interactions handled by AI agents in 2025, on the way to more than 34 billion by 2027. Growth on that scale marks the move from experiment to infrastructure.
From AI-Assisted to AI-Led
There is real distance between a seller who uses AI to write a sharper email and an agent that identifies a high-intent account on its own, sources the contact, launches a personalized sequence, and routes the reply to a human only when it is ready for a conversation. One makes a person faster. The other takes on the work.
This fundamentally changes what accountability looks like. When AI assists a seller, the seller owns the result. Gartner found that sellers who partner well with AI are 3.7 times more likely to hit quota, and that only 7 percent of sellers have strong AI-partnership skills today, which means most organizations are leaving that advantage on the table. When an agent is doing the executing, ownership gets harder to locate, and most revenue teams have not worked out what that means for how they measure performance, structure their teams, or govern the work.
The Gap No One is Measuring
The conversation with revenue leaders has shifted in a way I notice more every quarter. A year ago, the question was how to get sellers to use AI. Now it is how to know what the AI is doing, and whether it is working. Those are harder questions, and most of us are still building the muscle to answer them.
The reason is structural. Our performance infrastructure was built around people. Quota is assigned to a person, commission is paid to that person, pipeline reviews are built around what sellers did and said. So, when an agent sends 2,000 outbound touches, surfaces a qualified opportunity, and that deal closes two quarters later, what does your reporting actually tell you about the part the agent played? For most teams, very little, and that blind spot is expensive. Without visibility into what AI is actually doing in the revenue motion, you can’t optimize it, justify it, or scale it. That’s where the gap begins.
And it’s already opening. Gartner found this year that AI is saving sellers close to five hours a week, yet 72 percent of sales organizations are not reinvesting that time into higher-value work. One in four organizations reports a return of 50 percent or more on its AI investment. One in five reports the opposite, a negative return of the same size. The technology approaches tend to be similar, but what differs is the operating model around them.
Three Things to Do Now
This is an operational problem with operational answers. Here is where I focus.
First, treat an agent the way you would treat any participant in the revenue motion. When you deploy one, define what it owns, what it is measured on, and what a good outcome looks like. If an agent is running outbound, hold it to a reply rate, a meeting-conversion rate, and a pipeline number, the same way you hold a person to a number. That data must be placed where managers can see it, review it on a cadence, and act on it.
Second, build the governance before you need it. The teams that are ahead moved fast on adoption, and they also did the quieter work, writing down how AI operates inside the business, which workflows an agent runs on its own, which ones require a human to review, and what happens when something breaks. That documentation is what lets you stand behind agent-led execution in front of a customer, a board, or a compliance team.
Third, redefine what your managers are managing. As agents take on more of the execution, the manager’s job should grow with it. Pipeline inspection becomes less about asking a rep what they did and more about reviewing what an agent surfaced, flagged, and acted on, and coaching shifts toward judgment, escalation, and the handling of exceptions. Gartner is already advising sales leaders to prepare roles that orchestrate AI agents rather than simply use them.
The Real Question
So, no. The agent does not get commission. But commission is how we have always tied a result back to the person who earned it, and that’s exactly why the agent changes everything. It doesn’t replace the seller who earns it. It’s the unlock that lets them earn more: more coverage, more precision, more deals that actually close. The organizations that build their operating model around that reality will pull ahead. The ones that don’t will have to measure why they’re falling behind.
Simply put, teams that treat this as an operating opportunity will pull ahead. An agent needs a number it owns, a manager watching its work, and a clear answer to who is accountable when something breaks. That work is quieter than launching the next agent, but it’s what makes a forecast you can trust.
Nadia Rashid is Chief Revenue Officer at Outreach, the agentic AI platform for revenue teams. She has spent more than two decades in revenue leadership and go-to-market strategy, including senior roles at Seismic and Marketo, and has been part of four company exits.
