Warning: Your B2B buyers already know you. They're still asking AI who to call

AI is changing how B2B buyers choose vendors. Discover how GEO, open content, and credible proof can help New Zealand businesses get onto the AI-generated shortlist. 

Santosh Pandey, co-founder and CEO of Christchurch agency Ridiculous Digital, on why the relationship-driven South Island B2B playbook now has an AI gatekeeper: buyers - even the ones who know you - are asking the machine who's worth a call before they make contact.

Right. This is your next big client. They haven't emailed you, filled in a form, or downloaded the whitepaper your team sweated over for three weeks. Right now they're in front of ChatGPT, typing "suitable [your category] providers for a company [their requirements]" and reading the five names that come back. If yours is one of them, you're in with a shot. If it isn't - and God forbid if they were expecting you to be there - no bounce, no lost-form alert, nothing. That deal is gone now. To a competitor with a better GEO game.

And that, my fellow hardworking Kiwi marketers, is the B2B marketing manager's biggest nightmare. You did the hard yards, the right content, the right campaigns. But AI rained on your parade. And this is a big part of why the pipeline has felt impossible to prove lately.

I live in a relationship town – Christchurch. But most Aotearoa towns, Auckland included, can be quite relational. And this is happening even when the buyer knows you. The long lunch, the handshake history, the "we always use these guys" - none of it stops a buying group running the AI check first. The relational shortlist and the machine's shortlist are drifting apart, and only one of them is visible to you. Let’s try and talk about that in this piece.

The "Attrubution" Monday Meeting - a Weekly Nightmare

Right. You know the one. Pipeline review. Attribution questions. Marketing generated the leads, sales binned most of them, attribution re marketing dollars is harder than ever, and someone senior asks why. You know the work is working. You just can't show it in a spreadsheet - and that gap is where (especially Christchurch!) marketers go to lose arguments. 

Forrester surveyed close to 18,000 business buyers this year, and generative AI came out as the most meaningful source they use to research vendors - ahead of your website, your product experts, and your BD/sales team. Buyers use it to compare you against rivals, decide who's worth a call, even draft the internal business case. All before they show up anywhere you can measure. The early journey has not gone away, just that you do not get to be a part of it as much. So that mucks up the attribution, the pipeline filling process, the lot, if GEO does not get a seat at the table. And for the millionth time - it is NOT the same as SEO.

The Shortlist Is the New Game

We talked about this a wee bit at last year's Marketing South re clients having a few vendors already in mind before the process begins. One number explains what's at stake here for us all. 6sense looked at how B2B deals get won: 95% go to a vendor already on the buyer's shortlist when the formal evaluation begins. Read that again. 95%. The shortlist is the key. And if you have done your job right so far, you should likely be on that "traditional" shortlist.

Enter Copilot (or another AI tool). And so if an AI now helps write that shortlist and your name's missing, and that when it was probably expected, that is a MAJOR setback. There's no rejection email. It shows up as nothing - and nothing, repeated across a year, is a very expensive number that lands on nobody's dashboard.

It is not all doom and gloom, though: Forrester also found buyers treat the machine's answers as a starting point, pressure-testing them with people they trust. Humans still close deals (phew!). But that is a BIG risk and a major setback. What has materially changed is who writes the shortlist, and where, and what happens when we are not on it.

But Why Can't the Machine See You?

This is a complicated one, and there are many technical aspects to it - but let us give you a few bits. Think about where your best material lives. The long form document or whitepaper - gated behind a form. Pricing - "contact us". The proof of what you do - inside a PDF or a sales deck. An AI can't read any of it, won't fill in your form, and slides straight past "we empower your transformation" because there's nothing in it to use.

Adobe's traffic data makes it sharper: tech and software firms - your buyers - lead every industry in AI-driven visits, while big chunks of the sites Adobe measures come back unreadable to machines. The audience most likely to research you with AI is the one you're most likely invisible to. Fun times.

What does the machine reach for instead? The same thing a cautious buyer does - proof from someone other than you. Forrester counts nine outside influencers in a typical B2B decision - analysts, peers, review sites - the voices a buyer trusts precisely because they don't work in your marketing team. (Ridiculous clients will remember this interrogation during the discovery phase!) What others say about you now counts for more than anything you say about yourself. 

Think of it like getting onto an approved-supplier list. The machine runs the spec check before a human picks up the phone - can it read your site, are your claims specific and backed by numbers, do credible outsiders vouch for you. You pass, and you're on the list. You fail, and there's no conversation to lose, because there was never going to be one. Yikes!!!

What gets you on the list:

• Proof on open pages - case studies, integrations, pricing signals an AI can read (not a PDF, not behind a form)
• Numbers over slogans - "cuts onboarding from six weeks to nine days" is liftable; "trusted by leading brands" is noise
• The outside vote - G2 or Capterra reviews, Google Business reviews, analyst mentions, coverage somewhere the machine already trusts. Nine outside influencers per decision - be in their material

   

What it can't see, or won't reward:

• Gated content and "contact us" pricing

• Adjective-heavy positioning with nothing to quote

• Your best work, locked in a deck on someone's laptop

   

Fresh numbers corroborate this. LinkedIn's B2B Institute, working with Bain, found 40% of B2B deals collapse less because of a rival, but more because the buying group can't defend the choice. (Yes, LinkedIn sells ads to B2B marketers – our own filter applied; the Bain partnership and the method hold up.) For a Christchurch firm, that changes the size game: a case study from a business that looks like your buyer, beats a global major brand. Nice!

Measure What You Can't See
We love our dashboards and, as much as we'd like, we won't dashboard our way out of a dark funnel with one more tool. What the clever teams do instead is old-school: they lead with strategy and they ask. A "how did you hear about us?" box on every demo and contact form - free text, never a dropdown, because a dropdown only returns the options you already guessed. Free text catches the Slack thread, the podcast, the "ChatGPT suggested you". Then run the answers through AI to sort them at scale - a backend job, done quietly.

Treat what buyers tell you as a cross-check. Because they themselves may not remember the display ad they saw or the LinkedIn post they scrolled past. So still lead with strategy - this battle is fought on all frontiers at once. And run the buyer's question yourself: open ChatGPT or Copilot and ask who a [your category] buyer should call. Five minutes tells you plenty, and tracking how often - and how accurately - the machine names you now sits alongside branded search and that "how did you hear about us" box in the measurement kit. And when you report it, hand your CEO a range that is honest about what can and can't be known. A single confident number you cannot back gets pulled apart in that Monday meeting. A range is more resilient.

The Silver Lining - Again
The AI your buyers are asking was built and runs, thankfully, and so far, offshore. It has read most of the internet and barely heard of us down here. Ask it about your category and it can still show some of the big overseas names, mention a Christchurch firm in vague, half-right terms, or skip us altogether. This is us South Islanders' chance to beat the Goliaths of the world. This is how we level the playing field with both Auckland and New York!

Because the machine can't be outspent, and it can't be schmoozed. An Ōtautahi Christchurch consultancy with proof on open pages and reviews it didn't write can get named ahead of a national player whose best material lives in a locked PDF - and that national player can get ahead of the London one with worse hygiene. Same story as last edition. The field levels for whoever does the unglamorous work first.

The deals you pitch for and lose - you can see those, and learn from them. The ones that never reach you, because the machine never said your name - those you can't even count. The shortlist is being written right now, by something you'll never meet, for buyers you can't see. You can't sit in that room. You can make sure your name's on the list when they read it out. Your move is next!

Related reading:

If you sell products online rather than services, the same shift is rewriting your product pages too - we've covered what AI shoppers reward and ignore in our companion piece, "The second shopper". 

Wondering what the machines say when someone asks about your category - and whether your name comes up at all? That's a Ridiculous special. Bring your pipeline numbers and your best case study, and we'll show you what the AI sees when it goes looking for you.

Santosh Pandey is the co-founder and CEO of Ridiculous Digital, a Christchurch agency helping New Zealand businesses with Generative Engine Optimisation (GEO), digital strategy and execution, and paid media.