Recently I was working on a large deal with a global mining company. The buyer's concern was not our product. It was the future: what else could be coming with new AI agents, and how quickly things were changing.

No deck answered that. No demo answered it either.

What answered it was one morning, face to face. After that session, they knew we were the right partner.

That experience captures what I see across enterprise sales right now. AI is making the buyer more informed and, at the same time, more cautious.

Both sides of the table now have AI

Buyers and sellers are adopting AI at the same time. Forrester's 2026 research, cited by Gong, found that 94% of business buyers use AI in their purchase process, up from 89% the year before. On the seller side, Salesforce's State of Sales report found 87% of organizations use some form of AI and 54% already use AI agents.

The next step is agents talking to agents. Forrester expects about one in five B2B sellers to face agent-led quote negotiations by the end of 2026. Gartner predicts AI agents will intermediate 90% of B2B buying by 2028. That is a forecast, and intermediated does not mean decided.

What AI does well: it makes decisions faster

AI is excellent at the work around the decision. Buyers use it to research vendors, compare options and build shortlists. Sellers use it to prepare for meetings, personalize outreach and handle routine quoting.

The results are measurable. Gartner found that sales organizations giving sellers AI-enabled next best actions are 2.6 times more likely to achieve commercial growth. Gartner also predicts that by 2027, 95% of seller research workflows will begin with AI.

On the buyer side, procurement teams are already using agents to negotiate routine supplier terms inside guardrails they define in advance. Honeywell, for example, reports more than $500 million in spend and 2,500 negotiations handled through Pactum's agents. That is a vendor claim, but it shows the direction.

The pattern is consistent: AI shortens the time between question and answer. It does not shorten the time it takes for someone to decide they can live with the consequences.

Why trust lags behind adoption

Buyers use AI heavily and trust it sparingly. In the TrustRadius 2026 report, 94% of B2B buyers who used AI fact-check its output, and only 2% say they trust it always. Gartner found that 67% of buyers prefer a rep-free experience, yet 69% still validate AI-generated insights with a sales rep.

Three reasons stand out:

  • Verification burden. Forrester found that 19% of business buyers were less confident in a purchase because AI gave them inaccurate or misleading results.

  • No one to hold responsible. Forrester even predicts that a Fortune 500 company will sue a B2B provider in 2026 over AI-generated misrepresentation.

  • Doubt about whose side the AI is on. In a consumer survey, 75% said they would trust AI agents less if brand money swayed their recommendations.

Humans are not trusted blindly either. Gartner found that 51% of buyers expect misleading information from GenAI and 49% expect it from a sales rep. So the human advantage is not honesty. It is understanding and accountability: buyers were 39 percentage points more likely to say a human rep understood their needs than they were for generative AI.

The accountability test

In a large enterprise deal, every decision eventually reaches three questions. Who is putting their name on this? Who explains it to the board if it goes wrong? Who fixes it when it does?

An agent can produce the shortlist, the comparison and the first draft of the contract. It cannot answer those questions, because it cannot be held personally responsible. A system can be updated, switched off or replaced. A person cannot.

This is true on both sides. The buyer carries the risk of the decision, and the seller carries the commitment to deliver. Each needs to know the person across the table will still be accountable when the situation changes.

That was the gap in my mining deal. The buyer's question was about the road ahead: would we still be the right partner as AI and everything around it changed? Only people can make that kind of commitment to each other.

What sellers and buyers should do

For sellers

  1. Be easy for a buyer's agent to verify. Publish clear specs, transparent pricing and a current security and compliance trust center. Agents verify accurate data, clean transaction history and transparent terms.

  2. Lead with proof, not claims. Buyers value verifiable evidence of outcomes over assertions of market leadership. References and trials carry weight.

  3. Start from what the buyer's AI already told them. Ask what they found, then fill the gaps and go beyond the summary. Gong calls buyers who mention their AI research a gift.

  4. Automate the routine, keep humans on the exceptions. One automated seller can repeat a mistake across thousands of deals. Keep people on the judgment calls.

  5. Build relationships before you need them. Multi-thread across the buying group, show up for the validation moments, and invest in time together when nothing is for sale.

For buyers

  1. Delegate research and routine negotiation. This is where agents add the most capacity.

  2. Govern your agents. Decide what they may evaluate, on what criteria, and who is liable when they are wrong.

  3. Keep a named human accountable for strategic purchases, and meet the people you will depend on for years.

The deal gets faster. The accountability does not move.

AI will help buyers and sellers make decisions faster. Research that took weeks will take hours, and routine negotiations will run while we sleep.

But speed was never the hard part of a large enterprise deal. Accountability is. When the decision carries real risk, a person on each side still has to stand behind it, and people trust the people they know.

So as the deal gets faster, the relationship has to start earlier. AI can prepare the conversation. It cannot take responsibility for the outcome.

My mining customer did not need another demo. They needed a morning together to see who they would be working with when the world changed. In large enterprise deals, that is still where trust is built, and it is still our job.

Sources

Searched November 2026. Vendor-sourced figures are labeled in the text; Gartner and Forrester figures are as reported by the outlets below.