Gartner Says AI Agents Will Outnumber Sellers 10 to 1. Buyers Already Said What They Want.
TL;DR. Gartner predicted in November 2025 that by 2028, AI agents will outnumber human sellers ten to one, and that fewer than 40% of sellers will say those agents made them more productive. Meanwhile buyers keep saying the same thing in Gartner's own surveys: they avoid irrelevant outreach, they do their own research with AI, and they want a person to check the answer before they commit. More agents sending more email is the wrong race. The winners will use AI for judgment before the send, not volume after it.
The prediction
In November 2025, Gartner made a prediction worth reading twice: "By 2028, AI agents will outnumber human sellers by tenfold, but less than 40% of sellers will report that AI agents have improved their productivity" (Gartner, November 2025).
The first half is the headline everyone shared. The second half is the story. Ten agents for every seller, and by Gartner's own forecast, most sellers won't feel more productive.
That's what happens when you add capacity to the wrong part of the job.
What buyers say
Gartner surveys buyers too. Line the findings up next to the prediction:
- They avoid irrelevant outreach. 73% of B2B buyers actively avoid suppliers who send irrelevant outreach (Gartner, June 2025).
- They'd rather research on their own. 67% prefer a rep-free buying experience, and 45% used AI during a recent purchase (Gartner, March 2026).
- They want a person to check the AI. 69% prefer to validate AI-generated insights with sales reps (Gartner, May 2026).
- Gartner's 2030 prediction points the same way. It predicts that by 2030, 75% of B2B buyers will prefer sales experiences that prioritize human interaction over AI (Gartner, August 2025).
Line those up. Buyers want to do the research themselves, they use AI to do it, they avoid anyone who wastes their time, and they want a human to confirm the answer before they commit.
Nothing in that list says "send me more email."
The buyer side of this story, how AI search changed the way buyers check vendors, is its own post: Your Buyer Asked ChatGPT Before They Answered Your Email.
What the first AI SDR wave taught
The first generation of AI SDR products took the most visible part of the SDR job, sending emails, and automated it at enormous volume.
It's worth hearing how that went from inside the category. In April 2025, Artisan's CEO told TechCrunch that first-generation AI SDRs "get a pretty low response rate" and have "relatively high churn," and said the company had "extremely bad hallucinations" when it first launched (TechCrunch, April 2025).
That's a founder describing his own category. The lesson isn't that AI can't do outbound. It's that automating the send without automating the judgment produces exactly what buyers told Gartner they avoid: more irrelevant email, faster.
The SDR job was never mostly about sending. It was about deciding who's worth writing to, figuring out why they'd care now, and saying something true and specific. Automate the sending alone and you get a faster version of the worst SDR on the team.
If you're weighing an AI SDR tool against a service, we compared the two in AI SDR Tools vs. AI Outbound Agency.
Why most sellers won't feel more productive
Go back to the second half of Gartner's prediction: fewer than 40% of sellers will say agents improved their productivity. Why would ten agents per seller not help?
Because an agent pointed at volume creates work for the person downstream of it.
Someone has to check the output. An agent that writes a thousand emails produces a thousand chances to say something wrong about a prospect. If nobody checks, the mistakes reach buyers. If someone does, that's the seller's afternoon.
Someone has to handle what comes back. More sends mean more replies, and a lot of them are "not interested," "wrong person," "remove me," out-of-office notices and the occasional angry note. Sorting that is real work, and it lands on people.
Someone has to clean up the record. Every automated touch writes to the CRM. When the agent's view of an account disagrees with the rep's, somebody has to reconcile it.
None of that shows up as productivity. It shows up as a seller who has more activity on the dashboard and less time for actual conversations. That's the gap Gartner is predicting, and it comes from where the agents are pointed, not from the agents themselves.
Where AI belongs in outbound
Use AI where it gives a seller more judgment, not more volume.
Before the send. Research every account and every person. Decide whether to write at all. Find the reason this account would care this month. This is the slowest part of the SDR job, and it's where AI is strongest. (What prospect research means to us.)
In the writing. Every message written for the person reading it, built on facts that trace to a source. If a buyer checks your email against an AI answer, and many will, it should hold up.
At the reply. Sort and route replies fast, with the full history attached, so a person can pick up the conversation the same day. The reply is where the buyer asked for a human. Give them one. (How we handle replies.)
Sending more isn't on the list. Volume was never the constraint that mattered. Relevance was.
Where AI doesn't belong
Just as important is where to keep it out.
Deciding to keep going after a no. "Not interested" and "remove me" end the conversation. An agent that reads a no as an objection to handle is how a brand ends up in a screenshot.
Answering an upset reply. A prospect who writes back annoyed is giving you one chance to be a decent company. That's a person's job.
Filling a gap with something plausible. If the research can't confirm a detail, the email shouldn't contain it. A confident, specific, wrong sentence about someone's company does more damage than a plain one.
Pretending to be a person. If AI did the research and the writing, say so plainly, once, somewhere in the thread. A buyer who finds out on their own feels tricked. A buyer who's told can judge the work on its merits, and good work holds up.
A test for any AI outbound program
Ask these about your own program, or any vendor's:
- Would you send this email if a person had to write it by hand? If not, AI isn't making it better. It's making more of it.
- Can every fact in it be traced to a source? A made-up detail costs more than no detail.
- What decides who doesn't get an email? A program that never says no is a volume program.
- Who reads the replies, and how fast?
- If a buyer pasted your email into ChatGPT and asked "is this accurate?", what would it say?
The bottom line
The agents are coming either way. Gartner thinks there will be ten of them for every seller. The sellers who win won't be the ones with the most agents sending. They'll be the ones whose every message is worth checking, and who have a person ready when the buyer wants to talk. (How we think AI outbound should work.)
Frequently asked questions
Will AI agents replace SDRs?
Gartner predicts AI agents will outnumber human sellers ten to one by 2028, but also that fewer than 40% of sellers will say those agents improved their productivity (Gartner, November 2025). The parts of the job AI does best, research and first drafts, are moving to agents. The conversation is not.
Do buyers mind AI-written outreach?
Buyers mind irrelevant outreach: 73% actively avoid suppliers who send it (Gartner, June 2025). AI that sends more generic email makes that worse. AI that makes each email specific and accurate makes it better.
Where should a team start with AI in outbound?
With research and account selection, before the send. That's where AI saves the most time and where it most improves what gets written.
Want to see researched-grade treatment on one of your accounts?
Pick an in-market account that's currently slated for the template. Send me the company name and I'll show you what the research pass produces. No charge, no obligation.
Book a 30 minute walkthrough