Everyone Sounds Certain Now. Who’s Actually Right? B2B Marketing in the Age of AI

Confidence is in short supply, and everyone’s asking for it.

If you work in B2B marketing, you’re probably somewhere in that chain. Your leadership team wants to know which investments will drive growth. You want to know if your agency’s recommendations will actually work. Your agency is trying to understand customers whose research and decisions it can only partly see.

Your buyers have the same problem. Picking a vendor means asking colleagues to put money, time and their own credibility on the line. They need enough confidence to act, and enough evidence to bring everyone else with them.

Now AI has entered the chat. Buyers use it to research vendors. Marketers use it to produce campaigns. And nobody’s quite sure which signals, sources and explanations deserve trust anymore.

AI has made sounding certain easy. Too easy. But it hasn’t made certainty easier to back up when the tough questions come.

That tension runs through Pretzl CEO Greg Salmon’s conversation with Joel Harrison on the “Trust and Influence in B2B” podcast. From lead scores to attribution, Greg questions the tools that make marketing feel more predictable than buying really is.

Are we reducing uncertainty, or just circulating reassurance and hoping repetition will pass for proof?

Congratulations, you slept brilliantly

Greg compares the comfort of a lead score to his Apple Watch telling him he got an 82 for sleep. You know the feeling. Apparently you’re well rested, although you’ve already forgotten why you walked into the kitchen.

A score turns a messy experience into a tidy answer. But as Greg puts it, it “isn’t necessarily the full picture.”

Lead scores work the same way. Someone downloads a report, joins a webinar, visits the pricing page. Their score climbs. Marketing has something to report, and sales has someone to call.

It all adds up to a very convincing copy of confidence.

But one contact’s enthusiasm tells you almost nothing about whether finance backs the spend, whether legal is comfortable with the risk, or whether an influential colleague would rather stick with the incumbent and is scheming against you the whole time. Those people need different evidence, in different formats, at different times. And some of their influence will never show up in your systems.

Drawing them into a bigger diagram doesn’t fix it either. Greg warns against building a “3D version of the linear path,” an elaborate model built on the same flawed assumptions.

A confident explanation of a decision you can’t see

Joel put the visibility problem to Greg directly:

“If much of the journey can’t be seen, what should GTM teams stop assuming about attribution, intent and time?”

Greg called attribution “possibly the greatest villain in this space.” First-touch and last-touch models hand out credit, but neither tells you everything that went into a closed deal. And as AI agents do more of the buyer’s research out of the vendor’s sight, the idea that marketers can fully observe or control a buying journey gets even shakier.

AI also raises the price of being wrong. If your picture of the customer is off, producing more content and automating more activity just spreads the mistake faster. Scale up work built on shaky assumptions and you get more material that underperforms, and it gets harder to work out what’s actually helping.

What do we know, and what have we guessed?

This is where Customer Science has to earn its name.

Pretzl’s Customer Science manifesto describes a discipline built on testable hypotheses, evidence and continuous learning. What matters is what happens when the evidence contradicts an explanation everyone liked.

Greg draws a useful line here. Knowing someone is the CFO is a fact. Knowing how much sway they have over this particular purchase is often an inference. They deserve different levels of confidence.

The same goes for buying groups. Mapping one at a bank doesn’t mean you know how decisions get made at the next one. It gives you somewhere to start asking questions.

That matters to buyers, too. They have to defend the investment, deal with objections and win over colleagues. Greg asks how marketers can help them make that case internally. Our own belief that a campaign worked doesn’t give them the evidence they need.

Give uncertainty somewhere to go

Doubting everything can become its own excuse for doing nothing.

The Buyer Group Index is a practical place to start. Pretzl developed it with the ANA, with research by NewtonX, and it looks at 28 practices involved in buyer group marketing. The assessment shows where you’re strong and where you could improve.

Yes, it produces a score. After the sleep tracker, we know that’s worth a raised eyebrow.

But the useful part is what’s behind the number. Which capabilities can you actually demonstrate? Which answers are really hopes? Where would sales, marketing and data people disagree? Used that way, the assessment gives you priorities to investigate. It makes a huge subject manageable without pretending a number settles anything.

Trust what happens next

The confidence worth building is confidence in what happens when you’re wrong.

So watch what happens when the evidence disappoints. Does the team investigate, change the plan and say what it got wrong? Can it tell you what would change its mind?

That’s confidence in a method, and in the people using it. A good method needs people who’ll challenge their own assumptions. A capable team needs ways to test its judgment.

Admitting uncertainty isn’t worth much on its own. “We don’t know yet” only becomes useful when it comes with a credible way to find out.

Listen to the full episode of “Trust and Influence in B2B” with Pretzl CEO Greg Salmon and Joel Harrison.