The Seller's Edge: What Stays Human in an Agentic GTM World

A Meerkat and Sandler field report, adapted from the live fireside chat with Mary Shea, PhD, and Seth Marrs at the HSE AI Summit 2026.

Agent-to-agent GTM is coming, and in some scenarios it’s already here. But it won’t disintermediate the seller. It’ll do the opposite. In a market where machines handle everything that’s rote, the human seller becomes the edge, and that edge is about to become the revenue team’s most valuable asset. Beneath all the hype, the fundamentals of great selling haven’t changed nearly as much as people think.

Agent-to-agent is real, but it’s not universal

Agent-to-agent selling makes the most sense in high-volume, transactional sales where the product is well understood, the buying criteria are clear, and speed matters more than relationships. In those situations, agents can handle much of the work more efficiently, often delivering a better buyer experience. But big enterprise sales are where that logic starts to break down. These aren’t commodity purchases. They’re high-stakes decisions with real business and career consequences.

Agents will do a lot of the heavy lifting early in the buying journey. They’ll research vendors, compare solutions, and accelerate the evaluation. But when it’s time to make the decision and sign the contract, buyers will still want a human involved. Not because the technology isn’t capable, but because enterprise buyers aren’t just choosing a solution. They’re choosing the person they’ll call when things don’t go according to plan. In enterprise sales, buyers don’t just buy products. They buy accountability.

The lines won’t be drawn by what the technology can do. They’ll be drawn by what buyers want. The buyer chooses the interface, and the seller’s job is to meet them where they are.

The human is the brain. The AI is the brawn

The agent’s only as good as the agent boss. The human is the expert. The human is the decider. We do the thinking and set the guardrails. The AI takes the repetitive work and the research off your plate so you can carry more accounts, deliver deeper insights, and oversee more agents. As Mary puts it:

I offload the research, sequencing, content creation, and first drafts. I don’t offload judgment. I tailor everything. I beat the crap out of my AI agents before anything ships because AI-wonk language and generic content are relationship killers.

The result isn’t a diminished seller. It’s the best version of that seller. After nearly 30 years in sales, Mary has never been a better seller:

I can walk into a first meeting with the kind of executive brief that used to take a consulting team days to produce. I can role-play the toughest C-suite conversation before it ever happens. And I can combine decades of experience with research and insights that would never make it into the CRM I don’t own.

Here’s the principle underneath all of it: just because you can doesn’t mean you should. AI lets you fire off an infinite amount of outreach, and all that’s produced is more spam, not more trust. Cold outreach converts around 2%. A warm introduction converts between 15–25%. The winners won’t use AI to do the wrong things faster; they’ll use it to get back to more human, more connected selling, at scale.

Sales training’s stickiest problems are finally solvable

As Forrester analysts, we covered sales training extensively, and for decades the same three problems refused to go away. Could you see whether the methodology you invested in was actually being used in the field? Could you connect that adoption to measurable business outcomes? Could you help sellers adapt to the buyer in front of them instead of relying on one-size-fits-all best practices? The technology evolved. The problems didn’t.

What’s changed is that AI can finally solve all three, but only with the right infrastructure behind it. Dumping a call transcript into an LLM to assess methodology adoption is doable, but it’s manual and clunky. It doesn’t scale beyond a five-person team, and most enablement leaders abandon it after two weeks. The real answer is infrastructure: capture every customer conversation across the sales organization, score selling skills against a consistent benchmark, store those scores over time, and correlate them with the metric that matters most: revenue per rep. Then, and only then, can you tie methodology adoption to win rates, quota attainment, and business outcomes.

Do that, and the relationship between a training provider and its customer fundamentally changes. The job isn’t to sell methodology and training modules. The job is to be a growth partner who says, “Let’s make your methodology visible. Let’s measure whether it’s working. And let’s improve the performance of every seller.”

Five moves for revenue leaders

1. Bet on the seller, not against them. In an agentic world, human judgment, trust, and the ability to read the room become more valuable, not less. Give your sellers AI teammates. Don’t replace the people who close your biggest deals.
2. Make relationships, not sequences, your system of record. The best tools don’t just automate outreach. They help sellers discover who already knows and trusts them, turning your organization’s relationship network into a competitive advantage that buyers welcome.
3. Build around your buyer, not your technology. Some buyers will want AI agents to do most of the work. Others will want a human involved from the beginning. The buyer chooses the interface. Your job is to meet them where they are.
4. Stop treating sales training as an act of faith. You measure every other investment by business outcomes. Sales training shouldn’t get a free pass. If you can’t see who’s adopting the methodology and connect it to revenue per seller, you’re leading like it’s the 20th century.
5. Build super sellers, not AI replacements. Sales teams will shrink, but that’s not the story. The story is what the best sellers become when they’re paired with AI teammates. They prepare better, learn faster, build stronger relationships, and create more value in every customer conversation. That’s the future of selling.

The either-or debate, agents replace humans or they don’t, is dull, and we need to move on. The value stories piling up are about making sellers better, not replacing them. The magic happens when humans and machines work in concert, each playing to their strengths.

Meerkat is the relationship intelligence layer for AI agents. Sandler is a global sales training and performance-improvement organization.