AI search & AI transformation

Do the Buyer’s Work, and Their Agent’s: How to Win on Value

By Published Updated

Diagram: seven numbered circles in a loop around the words Revenue you can prove, step 4 of 7 highlighted. Each circle carries a letter for its job: A measurement, B AI answers, C AI-native team, D proof and positioning.

Step 4 of 7 in The Definitive Guide to Marketing in the Age of AI (in Seven Steps). Originally published November 4, 2024. Rewritten for the AI era October 26, 2026.

Content wins on value when it does work the buyer would otherwise have to do: frame the problem, compare the options honestly, build the business case and lower the risk of saying yes. Now that buyers send AI to do the first pass, that work has to be readable by a machine and checkable by a person. Do both and you compete on value. Skip them and you compete on price.

A note on proof: Proof below predates generative AI; an AI-era result replaces it by mid-December.

How are B2B buyers using AI to buy?

To gather information first, and to decide later with people. Gartner’s survey of 645 B2B buyers, fielded in August and September 2025, found that 45% used generative AI in a recent purchase, mostly to gather information on vendors and products. A later Gartner presentation put GenAI use during purchases at 60% (Gartner; Gartner).

But buyers don’t fully trust the machine. In the same survey, 69% said they prefer to validate AI-generated insights with a sales rep. And 51% said they’re more likely to meet misleading information from GenAI (49% said the same of reps). So the journey now has two checkpoints: what the AI found, and what a person confirms.

Two checkpoints in the B2B buying journeyFour boxes in a row: AI first pass, 45% of buyers used GenAI; Shortlist built from facts it can match; Human check, 69% prefer to validate AI insights with a rep; Decision. Source: Gartner survey of 645 B2B buyers, 2025.AI firstpass: 45%used GenAIShortlistfrom factsit can matchHuman check:69% validatewith a repDecisionSource: Gartner survey of 645 B2B buyers, Aug–Sep 2025.Two checkpoints in the B2B buying journeyFour boxes in a row: AI first pass, 45% of buyers used GenAI; Shortlist built from facts it can match; Human check, 69% prefer to validate AI insights with a rep; Decision. Source: Gartner survey of 645 B2B buyers, 2025.AI first pass: 45% used GenAIShortlist from facts it can matchHuman check: 69% validate with arepDecisionSource: Gartner survey of 645 B2B buyers,Aug–Sep 2025.
Buyers let AI do the first pass, then check it with people (Gartner, 2026).

What work is the buyer trying to get done?

Five jobs, roughly in order. Your content should do each one better than the buyer could alone.

The buyer’s five jobs, and what content does for each
The buyer’s job Content that does it What an AI agent needs from it
Understand the problem A plain explanation of causes and costs, with numbers Definitions and figures stated in one sentence each
Compare the options An honest comparison, including when you’re not the right fit A table with the same criteria for every option
Build the case A business-case outline with the inputs a CFO will ask for The inputs listed plainly: costs, timeline, team, risks
Lower the risk Proof from similar companies, implementation steps, what can go wrong Named, dated results; a step list; stated constraints
Get sign-off A short summary the champion can forward A standalone summary paragraph at the top

The last row matters most. Your real reader is often the champion who has to defend the decision in a meeting you’ll never attend. Arm that person.

Why do buyers still call a rep, and what should the rep carry?

Because AI can’t take responsibility for a decision, and a buying group needs someone who can. That 69% is good news for sellers who show up with proof. It’s bad news for those who show up with a pitch.

Build validation assets for reps:

  • A one-page “what AI gets wrong about us” that corrects common errors with sources.
  • The comparison table from your site, so the rep and the AI tell the same story.
  • Three dated, named results the rep can cite from memory.
  • A risk list: what can go wrong, and how you handle it.

How does content defend price?

By making value easier to see than cost. When buyers can’t tell options apart, price is the only thing left to compare.

I’ve seen this work. In 2020, during my D Custom years, McKesson’s new pitch reduced price pressure and increased its win rate. At D Custom (2014–2024), thought leadership written for generalists rather than specialists helped DDN shorten purchase cycles and win more deals.

What makes content agent-readable?

Clear facts, stated once, in the same words everywhere. An agent matching your company against a buyer’s instructions needs:

  1. A one-sentence description of what you do, for whom and where.
  2. Structured comparisons in HTML tables, not images of tables.
  3. Constraints stated plainly, such as minimum project size, regions served or what you don’t do.
  4. Proof with names and dates, linked to a page that holds the detail.
  5. Consistent entity details, so the agent doesn’t confuse you with someone else (see chapter 1).

This also helps industry pages. Specific beats general for both people and machines, which is why I recommend publishing industry-specific content.

Checklist: do the buyer’s work

  • One page per buyer job, or one page that does all five clearly
  • An honest comparison table, with the same criteria for every option
  • A business-case outline with the inputs a CFO asks for
  • A forwardable summary paragraph at the top of every key page
  • Validation assets for reps, including “what AI gets wrong about us”
  • Plain-text facts and constraints an agent can match

Where to go next

Doing the buyer’s work takes clear writing that people and machines can both use. Chapter 5 covers how to write for the skimmer, the scanner and the machine, and how an AI-native team produces it. If you want to see what AI tells your buyers about you today, and what your content is worth to the pipeline, ask about the measurement and AI-visibility audit: work with me.

Buyers don’t pay more for nicer words. They pay more when you’ve already done half their job.

Frequently asked questions

What does “win on value” mean in B2B marketing?

It means giving buyers enough clarity about outcomes, fit and risk that price stops being the deciding factor. Content does that by doing the buyer’s research and business-case work for them.

How do B2B buyers use AI in purchasing?

Gartner found 45% of B2B buyers used generative AI in a recent purchase, mainly to gather vendor and product information, and 69% prefer to validate AI-generated insights with a sales rep.

What content should sales reps have in the AI era?

Validation assets: a page correcting what AI gets wrong about you, a comparison table that matches your site, three dated and named results, and a plain risk list.

How do I make my content useful to AI agents?

State facts plainly and consistently: what you do, for whom and where, your constraints, structured comparisons in real tables and named, dated proof.

Abstract illustration: on the left, a teal ring with a gap at the top marked by a dotted light-blue line, labeled Open loop; an arrow points right to a full ring of five arrowed steps around a light-blue center, labeled Closed loop.
Content that sells

Thought Leadership and Curriculum Fail the Same Way

In business, content is thought leadership. In education, it’s curriculum and companion materials. Both fail when the loop stays open, but only schools inherit a ready-made test.