Your Next Customer Will Ask an AI First: What NZ Small Businesses Should Do About It

As at mid-2026 the numbers are in: 51% of B2B software buyers start their research with an AI chatbot more often than with Google, 68% of Google searches end without a single click to any website, and 63% of technology buyers used AI in their most recent purchase. Your next customer will probably ask ChatGPT, Copilot or Google's AI before they ever find your website — if they visit it at all. Here's what's verifiably true, what it means for an NZ small business, and the practical response.

The Numbers (All Sourced, All Dated)

We've kept every figure attributable, because a post about AI-era credibility should practise it:

Buyers moved first. Forrester reported in January 2026 that 94% of business buyers used AI in their buying process, and that buyers named generative AI or conversational search as a meaningful information source twice as often as any other source. G2's "Answer Economy" research (surveyed March 2026) found 51% now start with an AI chatbot more often than Google — up from 29% a year earlier — 69% ended up choosing a different vendor than they'd planned, and 33% bought from a vendor they'd never heard of before the AI mentioned it. That last number is the opportunity hiding in the disruption: challengers get discovered now.

Clicks are disappearing. SparkToro's analysis of Similarweb clickstream data (June 2026) found 68% of Google searches ended without a click to any website — up from about 49% in 2019. Ahrefs' February 2026 update across 300,000 keywords found that when an AI Overview appears, the top-ranking page's click-through rate drops 58%.

Trust hasn't moved as fast as usage. TrustRadius's 2026 B2B Buying Disconnect (published 15 July 2026) found 63% of buyers used AI in their journey — but 94% of them fact-check its answers at least some of the time. Gartner's May 2026 survey found 69% of B2B buyers prefer to validate AI-generated insights with a human sales rep, and G2 found 64% of buyers encounter inaccurate AI recommendations often or very often. AI opens the shortlist; humans and verifiable facts still close it.

What This Means for Being Found

The old playbook — rank, get the click, convert on your site — now leaks at step two. The new requirement is blunter: your information has to be good enough, and structured enough, to be the answer. For an NZ SMB, the practical version:

1. Answer real questions directly on your pages. Lead with the answer, then explain. AI systems extract passages, not pages — a clear 50-word answer to "how much does X cost in NZ?" is worth more than a thousand words of positioning.
2. Date everything and keep it true. "As at July 2026" beats undated claims — for AI systems weighing freshness, and for the 94% of buyers who fact-check. Stale pricing on your site is now actively harmful.
3. Publish specifics competitors won't. Prices, timelines, honest limitations. AI systems disproportionately cite content with concrete numbers and sourced claims — and buyers arriving pre-informed convert with less friction.
4. Let AI crawlers in. If your robots.txt blocks GPTBot, PerplexityBot, ClaudeBot or Google-Extended, those systems can't cite you. Blocking them is a legitimate choice — but not if AI visibility is the goal.
5. Mind your third-party footprint. AI answers lean on reviews, directories and independent coverage, not just your site. Accurate Google Business Profile, real reviews, consistent details everywhere.

We practise this on our own site — named authors on every post, FAQ and article structured data, an llms.txt file describing our services to AI systems, AI crawlers explicitly allowed, and published pricing with dates. None of it is exotic; all of it is deliberate.

What This Means for Selling and Serving

The same shift is happening on your side of the counter. Salesforce's Agentforce Commerce, GA since late June, can now surface product catalogues inside ChatGPT, with Google AI Mode integrations rolling out — the storefront is going wherever the asking happens. On service, Gartner predicts agentic AI will autonomously resolve 80% of common customer service issues by 2029, and outcome-based pricing — US$2 per resolution on Salesforce's new Help Agent — has made piloting cheaper to justify than ever.

But hold the counterweight stats close. Gartner also predicts over 40% of agentic AI projects will be cancelled by end-2027 — escalating costs, unclear value, inadequate risk controls. And in One NZ's AI Trust Report (May 2026, n=1,001), 62% of New Zealanders said they'd stop using a product or service over concerns about how an organisation uses AI, while 68% said the availability of human contact would increase their comfort. Translation: deploy AI where it demonstrably helps, keep humans reachable, and be transparent about which is which. Our Agentforce guide covers the sensible starting shape: one agent, one job, measured.

The NZ Reality Check

The local data says most of your competitors haven't sorted this yet:

• MYOB's Business Monitor (July 2025): 32% of NZ SME decision-makers use AI — up from 17% a year earlier, but still a minority.
• Datacom's State of AI Index 2025: 87% of NZ organisations use some form of AI, yet only 12% have scaled it across the organisation.
• 2degrees/Deloitte research (April 2026): the average AI-adopting NZ SME earned roughly $400,000 more in FY25 than a comparable non-adopter — the first NZ-specific estimate linking adoption to firm-level results.
• CPA Australia's 2025–26 Asia-Pacific Small Business Survey ranks NZ small businesses lowest of 11 surveyed markets on nearly all technology-adoption indicators.

Read together: adoption is broad but shallow, the laggard cost is now quantified, and the bar for standing out — in AI answers and in AI-assisted operations — is lower in NZ than almost anywhere. That's an advantage for whoever moves deliberately first.

What We'd Do This Quarter

1. Ask the machines about yourself. Put your five most valuable customer questions to ChatGPT, Copilot and Google. Are you mentioned? Is what's said accurate? That's your baseline.
2. Fix the answers you control. Publish direct, dated answers to those five questions on your site. Update anything stale — especially pricing.
3. Write down what only your team knows. The undocumented answers in your people's heads are invisible to AI search and unusable by AI agents. Documenting them serves both at once.
4. Get the CRM and knowledge base clean. Every AI initiative — visibility or agents — stands on this. It's the least glamorous item and the highest-leverage one.
5. Pilot one agent on one job. Documented questions, clean data, measured outcomes, human escalation. Expand only on evidence.

Steps 3–5 are exactly what our free Salesforce health check assesses — data, knowledge and readiness — and where our Agentforce work picks up if the pilot makes sense.

Frequently Asked Questions

Do customers really use AI to research businesses now? Yes, measurably. G2's March 2026 survey found 51% of B2B software buyers start research with an AI chatbot more often than with Google, and Forrester reports 94% of business buyers used AI somewhere in their buying process. TrustRadius's July 2026 report found 63% of technology buyers used AI during their purchase journey.

What is zero-click search and why does it matter for my business? Searches that end without the user visiting any website — because the answer appeared on the results page or in an AI summary. SparkToro's analysis of 2026 clickstream data puts it at 68% of Google searches. It matters because being technically "ranked" no longer guarantees being seen: your information has to be good enough to be the answer itself.

How does a small business show up in AI answers? Publish direct, factual answers to the questions customers actually ask — with dates, specifics and named authorship. Keep information current, allow AI crawlers in robots.txt, use FAQ and article structured data, and maintain accurate third-party listings and reviews. AI systems favour content they can verify and extract cleanly.

Should an NZ small business be using AI agents yet? For narrow, well-defined jobs on clean data — answering documented customer questions, qualifying leads — yes, pilots are cheap and the tooling is mature. For everything at once, no: Gartner predicts over 40% of agentic AI projects will be cancelled by end-2027, mostly from unclear value. Start with one agent, one job, measured.

What does CRM data quality have to do with AI visibility? Everything downstream. The same discipline that makes your website citable — accurate, current, structured information — is what makes AI agents on your side work: agents answer from your CRM and knowledge base. Clean data and written-down knowledge are now marketing assets, not just admin hygiene.

---

SaaSKool helps NZ and AU SMBs get their data, knowledge and Salesforce org ready for the AI-first era — Agentforce design and deployment, integration, and the groundwork underneath it all. Start with the free health check.