Skip to content

Analysis · Revenue and commercial operations

You don't create demand. You find out where it's already heading.

In 1966, Eugene Schwartz wrote that advertising cannot create desire. B2B buyer research from 2026 says the same, now with AI in the middle of the journey.

By the time a customer reaches out, the choice is mostly made.

B2B buyers surveyed in 2025, in %

Already had a preferred vendor before first contact

94%

Prefer to buy without talking to a sales rep

67%

Used generative AI in their latest purchase

45%
Source: 6sense, 2025 Buyer Experience Report; Gartner, survey of 646 B2B buyers, August and September 2025.

When sales stop growing, the usual diagnosis is that not enough people know the company exists. The fix comes as more ads, more sales reps, a new website and, lately, an AI agent answering on WhatsApp. Behind all of it sits one bet: that you can create the demand that’s missing.

Eugene Schwartz bet the other way. In Breakthrough Advertising, published in 1966, he wrote that “copy cannot create desire for a product.” The desire is already out there, shared by thousands of people at once, before any campaign runs. What a company can do is find that desire and give it a path to what it sells. Schwartz called it channeling.

Sixty years later, the research on how businesses buy backs him up. The difference is that part of that path now runs through artificial intelligence.

What the 2026 research shows

Gartner surveyed 646 B2B buyers in August and September 2025. In the results released in March 2026, 67% said they prefer to buy without talking to a sales rep, and 45% used generative AI in their latest purchase. In the findings released in May, buyers said they used seven sources of information on average, and 69% turn to a sales rep to validate what the AI told them.

6sense, which studies how businesses buy, put it more bluntly in its 2025 report: 94% of buying groups have a preferred vendor before they first talk to any seller. That favorite wins the deal roughly 80% of the time.

Put the two together. When a customer calls, asks for a quote or sends a message, they’ve already done the research without you and already have a shortlist. Almost always, they already have a favorite.

Advertising doesn't create the desire. It steps into a path the desire already takes.

From desire to purchase, in a business-to-business sale

  1. Desire

    Exists before any ad runs

  2. Research

    Search, AI, peers and reviews

  3. Shortlist

    The favorite comes from here

  4. First contact

    To confirm the choice

  5. Purchase

Where a company can still get in

The path shows the problem with trying to create demand. Ads and sales reps usually show up at step four, after the shortlist already exists. At step three, what counts is what the buyer found on their own: the search result, the AI summary, the story a peer told them.

In practice, channeling means being there at step two with the information that helps the buyer build the shortlist. That includes what an AI assistant finds about your company when someone asks which vendors to consider. The AI builds that answer from what’s already published: the website, the reviews, the case studies and what customers write.

There’s a second place where demand gets lost, and it’s inside the company. It’s the request that waits days for an answer, the quote that depends on one person, the message nobody answers over the weekend. If the customer arrived with a favorite on the list, every hour of waiting raises the odds they close with the favorite. In that case, more advertising just makes the line longer.

What I’d look at first

Before signing off on budget to generate demand, I’d put two questions side by side. Are people already looking for what the company sells? And can the company properly serve the ones who already show up?

Before paying to create demand, find out which quadrant the company is in.

Existing demand and the capacity to serve the customers who already show up

  • Fix operations first.
  • Channel the demand.
  • Rethink the offer.
  • Test before you advertise.

The answers usually sit in places the company already has. Messages and emails show what customers ask for without anyone offering it. Search Console and Google Trends show how many people look for what you sell, and in what words. And the time between a request coming in and getting a first reply shows whether the company can keep up with the customers already at the door.

With that in hand, the call gets simpler. If demand exists and gets lost inside the company, the money goes to operations first. If it exists and the company can keep up, the job is to channel it: show up where the customer researches, with the answer they’re looking for. If demand is thin and the company has spare capacity, test with a small budget before advertising at scale. And if demand barely shows up and the company still serves poorly, the conversation goes back to the offer, before any budget.

What we still don’t know

Neither survey breaks out Brazil, and I haven’t found comparable numbers for mid-sized Brazilian companies. It’s also unclear how much weight AI will carry in the shortlist. In 2025, nearly half of buyers were already using generative AI, but without much trust: 51% said they’re more likely to run into misleading information from generative AI, and 49% said the same about sales reps.

Do this tomorrow

Pull the messages and emails from the last 30 days that asked for a price or a quote. Count how many came in without anyone at the company pitching anything, and log how long each one waited for the first reply.