AI tools are rewriting the B2B buying process in real time

Amid the hype and dire warnings about AI creating a “Frankenstein’s monster,” one of the most tangible shifts marketers face today is the rapid transformation of B2B buying behavior. Deep research tools like OpenAI and Perplexity allow buyers to circumvent traditional processes, streamlining decision-making in ways that render old sales funnels obsolete.

TrustInsight.AI co-founder and chief data scientist Chris Penn provided a clear example of this shift. 

When a SaaS vendor raised prices, he asked Gemini Deep Research to identify five alternative providers, ranked by price and fit. “In 15 minutes, it gave me a list of five new vendors, all of which had price decreases compared to the vendor I was working with, and I switched to a vendor off the list.”

He cut the infrastructure costs in half while doubling service output. The original vendor never knew why they’d lost the business. “The new company also has no idea why, because I just went onto their self-serve signup portal and swiped the credit card, and now I’m using their services very happily.” 

The new reality

This is the new reality: instantaneous, AI-driven decision-making. It doesn’t download white papers or give you any clue who wants to buy, and it bypasses intermediaries like sales reps.

“That’s a big change because when you think about it from a CMO perspective, they’re like, ‘How are we attracting and retaining customers?'” he said. “It fundamentally challenges how complex sales are handled.”

Dig deeper: How omni-conscious brands win in a buyer-first, fragmented world

If the buyer does talk to a rep, it’s a different type of conversation. B2B buyers will come armed with AI-assisted evaluations of all the vendors, including pricing, service agreements, etc. 

AI can compress buying cycles dramatically for larger companies with complex, committee-driven purchasing processes. Stakeholders can rely on AI-generated shortlists built around specified criteria, shifting the onus to vendors to maintain explicit, searchable and accessible content — especially pricing — on their websites.

No more ‘registration required’

In this landscape, companies that hide pricing behind “call us” walls won’t make it onto AI-generated shortlists. The AI isn’t going to pick up a phone and wait for sales reps to respond. A lack of transparency becomes a disqualifier — not a negotiation tactic — and can result in missed revenue opportunities.

“If I said I want a new CRM, and the cost has to be $30 a month per user per seat or less, and you don’t have pricing on your website, guess what? You don’t make the list anymore,” said Penn. “This means traditional search optimizations are still very important because these AI tools are all grounded in traditional search.”

Navigating truth and trust in the AI age

While AI enhances research efficiency and purchasing speed, it raises new concerns about accuracy, user understanding and ethical design. Savvy users are learning to assess the reliability of AI outputs based on what’s at stake. 

“If it’s ‘Hey, I want a new Canva replacement,’ you’ll fact-check a little bit,” said Penn. On the other hand, if it’s, ‘Hey, I’ve got some medication I need to check,’ you’ve got to fact-check the heck out of that because if you don’t, you might end up hurting yourself.” 

Dig deeper: Why marketing must reclaim GTM design in the age of AI

It’s essential to remember that AI systems are designed to be helpful first, harmless second and truthful third. This means they often behave like digital “yes men,” producing agreeable — but not necessarily accurate — responses. This leads to bad decisions in business and can be even worse for individuals.

“One of the co-investors of OpenAI is having what’s being called ChatGPT psychosis,” said Penn. “He is publishing delusional rants on his social media based on stuff that ChatGPT is spitting back to him, which is derived from fan fiction. This user is getting answers that the tool thinks are helpful, even though it’s not being truthful, and it’s made this person literally have mental health issues.”

Don’t trust AI like you trust search engines

People have learned to trust search engines because they deliver accurate results. In part, because if the top answer seemed off, you could see if the other results said the same thing. That’s a flawed way to validate the answers because Google or whoever may have only given results that said the same thing. However, at least the users are trying.

Now, GenAI is delivering one answer, and users trust it as they did search engines. That’s because users aren’t taught how to interrogate AI results effectively. Most do not know how to prompt for dissenting views, identify missing context or push back on potentially biased or incomplete outputs. 

“You have to say, ‘Hey, challenge my assertions. Challenge my assumptions. Tell me what I’ve gotten wrong. Tell me what I’ve missed. Hey, what could go wrong with this? What did we overlook? What did we miss?'” said Penn. “These are all questions you should ask in your prompting process. And the reason why this isn’t happening is that these incredibly powerful tools come with no manual. It’s like getting a chainsaw with no manual. Like, ‘Okay, well, where should I grab it?'”

Trust, but verify

B2B marketers must understand this. GenAI is becoming a larger part of marketing workflows, which means there are more places where it can provide an answer that seems helpful, but is wrong. As Soviet leader Leonid Brezhnev once said, “Trust, but verify.”