When a homeowner asks Perplexity “who’s the best tree surgeon near me?” or reads Google’s AI summary above the results, an answer engine has just made a decision on your behalf: name you, or name someone else. There is no page of ten links to climb — just a shortlist of two or three businesses and a reason for each. This guide explains the ranking logic underneath that decision, so a tree surgery business can understand exactly why it is chosen or skipped.
It is part of our wider hub on GEO & AI SEO for Tree Surgeons. If you want the practical to-do list rather than the theory, our companion piece on how to get recommended by ChatGPT as a tree surgeon walks through the steps; this page is about why those steps work.
How does an AI engine actually pick a local business?
An AI engine does not hold a stored opinion about your business and read it back. It builds a fresh answer every time someone asks, using a retrieval-and-ranking process. Understanding the stages is the whole game, because each stage is a checkpoint you either pass or fail.
Most answer engines follow a version of this pipeline:
- Understand the question. The engine parses intent and location — “tree surgeon”, “emergency”, “near [town or postcode]” — and decides it needs current, local information rather than general knowledge.
- Retrieve candidate sources. It gathers documents and data about local businesses: Google Business Profiles, directory listings, review pages, news and your own website. This is the step that draws on Bing’s index, Google Maps, or a live web search depending on the engine.
- Score and filter. Each candidate is ranked on relevance, trustworthiness, freshness and how clearly it states the facts. Weak, contradictory or stale sources are dropped.
- Generate the answer. From the survivors, the engine writes a short recommendation — usually naming the two or three businesses it is most confident about, with a one-line reason each.
The businesses that get named are simply the ones that survive every checkpoint with the least doubt attached. Your job is to leave no doubt.
Where does each engine look for local businesses?
The biggest practical difference between the engines is where they retrieve from. Optimise for only one and you will be invisible on the others.
| Answer engine | Primary local sources | What that means for a tree surgeon |
|---|---|---|
| ChatGPT (Search) | Microsoft Bing’s index plus OpenAI’s own crawler; leans on Google Business Profile data | Claim Bing Places — not just Google — or ChatGPT may never see you |
| Google Gemini | Grounds local answers directly in Google Maps / Google Business Profile | A complete, accurate Google Business Profile is non-negotiable |
| Perplexity | Searches the live web and favours credible, recent, well-cited pages | The authority and freshness of pages that mention you carry real weight |
| Google AI Overviews | Google’s own index and local data, shown above the classic results | Strong local SEO and clear, answer-first content feed the summary |
The common thread is that all four assemble their answer from your Google Business Profile, directory listings, reviews and website — then favour whichever business is clearest and best corroborated. The differences are about which copy of those facts each engine happens to read first.
Why do AI engines recommend so few businesses?
Because they are far more selective than Google’s results page. Google will happily list you even if you are the eighth-best option in town. An answer engine names only the businesses it can recommend confidently. 2026 research from SOCi’s Local Visibility Index found AI assistants surface only a sliver of local businesses for a given query — roughly 1% of locations on ChatGPT, around 7% on Perplexity and about 11% on Gemini. A list of ten becomes a shortlist of two or three.
That selectivity reframes the task. The engine is not asking “who is good enough to list?” It is asking “who can I name without being wrong?” For a tree surgeon, the answer comes from having no gaps and no contradictions in the facts an engine can find — because every gap is a reason to choose someone safer.
It also makes this first-mover territory. Most tree surgeons are not optimised for answer engines at all yet, so the firms that get the fundamentals right now build authority that compounds as more homeowners shift to asking assistants instead of scrolling Maps.
What signals decide the ranking?
Once an engine has retrieved its candidates, it scores them. These are the signals that move you up or down the shortlist, roughly in order of how much weight they carry for local trade queries.
| Signal | What the engine is checking | How a tree surgeon strengthens it |
|---|---|---|
| Consistency (NAP) | Do your name, address and phone match everywhere? | Identical details on Google, Bing, Checkatrade, Yell and the Arboricultural Association |
| Corroboration | Do independent sources confirm you exist and vouch for you? | Reviews, local news, supplier “find a trader” pages, association listings |
| Reviews | Volume, average rating, recency and replies | A steady trickle of fresh, replied-to reviews naming real jobs |
| Clarity | Does your site plainly state services and service areas? | A page per service and per town, written in plain language |
| Structure | Can a machine parse who you are and what you do? | LocalBusiness schema; answer-first paragraphs; question-based headings |
| Freshness | Is the information current and being updated? | Recent reviews, updated hours including storm cover, dated content |
How much do reviews really matter?
More than most owners assume — reviews behave like a filter, not just a tiebreaker. Studies of AI recommendations in 2026 found engines applying effective rating thresholds, tending to recommend only businesses above roughly four stars, with the bar varying slightly by platform. Recency and replies matter too: thirty reviews from three years ago reads as dormant, while a regular trickle (“turned up within hours after the storm and took down a split oak safely”) signals an active, trustworthy firm. Reviews are free corroboration the engines can read, so they reward the businesses generating them consistently.
Why does structured, answer-first content win citations?
Because engines reward pages they can lift from cleanly. Content with proper schema markup and clear question-based headings is markedly more likely to be cited in AI answers, because the engine can extract a self-contained fact and attribute it with confidence. A page that buries “we cover emergency tree removal across [town] and the surrounding postcodes” inside a wall of marketing copy is harder to quote than one that states it in a single clear sentence under a relevant heading. This is also where being a recognised entity pays off — our guide to entity SEO and topical authority for tree care explains how to make an engine certain who you are before it decides whether to name you.
What makes an AI engine not recommend you?
It is worth naming the failure modes directly, because they are the cheapest things to fix:
- Contradictory details. A different phone number on Yell than on Google, or an old address lingering on a directory, makes the engine unsure which “you” is real — so it picks a cleaner candidate.
- A thin or unclaimed profile. Missing categories, no services listed, no hours. A half-finished Google Business Profile is a half-finished answer.
- Stale signals. No recent reviews, no updated content. Freshness checks read this as a dormant business.
- Bing invisibility. Because ChatGPT runs on Bing’s index, ignoring Bing Places quietly removes you from one of the largest answer engines.
- No third-party corroboration. If only your own website says you are good, the engine has nothing to verify against — and verification is what lowers its risk in naming you.
Every one of these is an “I can’t tell” that pushes the engine towards a competitor. The work of GEO is turning each one into an unambiguous “yes”.
How do you turn the ranking logic into a plan?
You audit yourself the way an engine would. Run through this checklist and treat each unticked box as a reason an AI might currently skip you:
- Google Business Profile claimed, verified and 100% complete
- Bing Places listing claimed and matching Google exactly
- Name, address and phone identical across every directory
- Steady stream of recent reviews, all replied to
- A page for each service (removal, crown reduction, pruning, stump grinding, emergency call-outs) and each town you cover
- LocalBusiness schema declaring services, service areas and contact details
- Answer-first, question-led content for common homeowner and council queries
- Mentions on trusted local or trade sources that corroborate you
- A monthly habit of testing the engines and checking referrals
This is the same groundwork that underpins ordinary Local SEO for tree surgeons — answer engines simply raise the reward for doing it properly, because they are far less forgiving of gaps than Google’s list ever was. If you want it done as one joined-up programme, our local SEO service covers the profile, citation and review work that feeds both classic and AI search.
How do you know if any of this is working?
You measure it, rather than guess. AI engines increasingly show up in GA4 as referral traffic from domains like chatgpt.com and perplexity.ai, and you can segment those visits, watch for branded searches that rise after an AI mention, and pair the data with call and form-fill tracking to tie an enquiry back to its origin. This is exactly where our data and analytics background earns its keep — the same approach behind our Jax Tree Removal rebuild and lead-tracking setup, where we track every lead and prove which jobs came from which clicks rather than hoping the work landed.
If you want to see what ChatGPT, Gemini and Perplexity can currently find about your business — and which gaps are keeping you off the shortlist — request a free audit and we’ll show you exactly where you stand and which fixes will move you up first.
Sources: SOCi Local Visibility Index 2026 (AI assistant recommendation rates); Stackmatix, “Structured Data for AI Search” (2026); ZipTie, “How Perplexity AI Answers Work” (retrieval, ranking and citation pipeline).