Here's a question that's been nagging at me: "What happens when the buyer no longer needs you to figure out what they need?" That question sits at the center of the shift that we are experiencing. 

For years we built sales careers on the ability to run strong discovery. We asked better questions than the next guy. We uncovered pain the buyer hadn’t fully articulated. We shaped the requirements in our favor. 

 That game is ending!

Most sales processes still treat discovery as the foundation. You open the first real meeting with some version of: 

• “Tell me about your current situation,” 
• “What’s driving this initiative,” or 
• “What does success look like for you?” 

You dig. You listen. You reframe. You try to control the conversation early so the requirements end up looking a lot like what you sell. The assumption underneath is simple: the buyer is still forming their thinking and needs a skilled salesperson to help them get clear. That approach worked when buyers had limited data and limited time to analyze it. They needed us to structure the problem and offer up a solution.

In my book, 'Future of Selling: The Rise of AI Agents', I argue that buyer-side AI agents will change that sequence. How? The agent pulls:

• public information, 
• internal data the buyer feeds it, 
• competitor comparisons, 
• pricing patterns, 
• implementation notes, and 
• support history. 

It then scores vendors against a clear set of requirements the buyer already defined. Only the vendors that clear the minimum threshold get a meeting. By the time you sit down with the human buyer, the heavy lifting of discovery is already done. The buyer is not coming to explore, but rather to evaluate the short list.

You no longer get to run discovery in the traditional sense. The buyer controls the flow. You are only in the room because the agent already decided you meet the basic specs. From that point, the conversation narrows to three things the buyer still needs from a human seller. I call them the 3Cs.

Clarification — answers to the specific questions the agent could not resolve with the data it had. These are usually edge cases, integration realities, or organizational implications that never show up cleanly in public sources.

Confirmation — direct proof that your product or service actually does what the agent thinks it does under their conditions. They want to hear you say it and show it, not dance around it.

Confidence — the human judgment that reduces the remaining risk. This is where product depth, real market experience, and a clear explanation of how the first 60–90 days will actually work become decisive. The buyer wants to feel they can move forward without looking foolish later.

B2B enterprise sellers who continue walking into first conversations assuming the buyer still needs help defining the problem will waste time and lose credibility. The new reality requires a different opening posture. Before you start selling, you need to understand how much the buyer already knows. The first questions in the AI-agent era sound more like this:

1. “Tell me how much you know, and I’ll begin at the edge of your knowledge.” 
2. “How much research has your team already done, and what have you found?” 
3. “What did your agent surface that still needs clarification from our side?”

These questions immediately move the conversation to the only place a human seller can still add value: the residual uncertainty that remains after the agent finishes its analysis.

The Rise of AI Agents

AI agents are not coming. They are already here and scaling faster than most sales leaders are willing to admit! Look at the numbers. 

• Salesforce’s Agentic Enterprise Index (August 2026) shows the average number of activated agents per organization jumped from 5 in February 2025 to 13 by April 2026; nearly a 3X increase in just over a year. 
• Box’s 2026 State of AI in the Enterprise found that 83% of organizations are already running AI agents, with 19% operating them autonomously at scale. 
• KPMG’s Q2 2026 AI Pulse reported that employee adoption of AI agents more than doubled in a single quarter, rising from 23% to 56%. 
• Gartner projects that 40% of enterprise applications will embed task-specific AI agents by the end of 2026, up from less than 5% the year before.

These are not pilot experiments. Large enterprises are putting agents to work on research, vendor scoring, requirements mapping, and short-list creation right now. When a procurement team or a line-of-business buyer feeds their constraints into an agent, that agent does in hours what used to take a buying committee weeks. The vendors who do not clear the agent’s threshold never get the meeting.

The companies and sellers who start treating discovery as a verification and confidence-building exercise, rather than a need-finding exercise, will keep earning the meetings that matter. Those who keep asking the old questions will be filtered out before they ever get the chance.

Discovery, as most of us were trained to run it, is on its deathbed...gasping for air!

Victor Antonio

p.s., Consider helping me out by pre-ordering my new book, "Priced to Win: The Sales Leader's Framework for Closing Deals Without Caving on Price," due out in Nov. 2026.

#ModernSelling #B2BSales #skospeaker #AIAgents #SalesLeadership #DigitalTransformation #BuyerJourney #SalesTraining #FutureOfWork #ArtificialIntelligence #futureofselling