Ask an assistant for “a good conveyancer in Melbourne’s inner north” and it will name three or four. Whether you are one of them is what generative engine optimisation decides. This guide explains what GEO is, where it overlaps with the SEO you already do, which engines are worth your attention, and how to measure the result rather than take an agency’s word for it.
What generative engine optimisation actually means
Generative engine optimisation, usually shortened to GEO, is the work of making your business the source an AI answer engine draws on when it writes a reply. The engines that matter in Australia are Google’s AI Overviews and AI Mode, ChatGPT, Gemini, Perplexity and Microsoft Copilot. Each one retrieves a shortlist of web pages, reads them, and composes an answer that may name and link a handful of sources. GEO is about being on that shortlist and being worth quoting once you are.
The name is new; most of the work is not. The engines still find pages through search indexes, still favour pages that answer a question plainly, and still trust businesses whose details are consistent everywhere they appear. What changes is the unit of competition. In classic search you competed for a rank. In generative search you compete for a citation inside one synthesised answer, and most answers cite three to six sources at most.
You will also see the terms AI search optimisation, answer engine optimisation and LLM optimisation. They describe the same discipline. We use AI SEO for the service, and GEO when we mean the technique.
GEO vs SEO: what changes and what doesn't
The honest summary: the inputs overlap by perhaps three quarters, and the measurement is where they split. A page that ranks well on Google is far more likely to be cited by an engine than a page that does not, because the retrieval step leans on the same indexes and the same quality signals. Google has said that AI Overviews do not need special markup and that standard SEO practice applies. That has matched what we see on client sites.
| Area | Classic SEO | GEO |
|---|---|---|
| Goal | Rank in the top results | Be cited inside the answer |
| Unit of competition | A position on a results page | One of a few sources per answer |
| Page structure | Keyword-aligned headings | Answer stated first, then evidence, in extractable chunks |
| Trust signals | Links, reviews, brand mentions | The same, plus consistent entity facts across the web |
| Freshness | Matters for some queries | Matters more: engines prefer recently updated sources for time-sensitive answers |
| Measurement | Rankings, clicks, conversions | Citations, AI referrals, assisted conversions |
What does not change: technical health, crawlability, a clear site structure, real expertise on the page and a Google Business Profile that is complete and reviewed. If those are weak, GEO work on top of them is wasted.
Which engines matter for a Melbourne business
Not every engine deserves equal effort. Weight them by where your customers actually are.
- Google AI Overviews and AI Mode. The largest audience by far, because they sit inside Google search itself. For most Melbourne service businesses this is the one that moves the needle. Overviews appear mainly on informational and comparison queries, less often on short commercial queries like “plumber Richmond”, where the local pack still dominates.
- ChatGPT. The most-used standalone assistant. Its web search draws on a live index, and it cites sources with links. Referrals from it show up in GA4 as traffic from chatgpt.com.
- Gemini. Google’s assistant, increasingly woven into Android, Workspace and Chrome. It leans on Google’s index and on Business Profile data for local answers.
- Perplexity. Smaller audience, heavy on citations, popular with researchers and B2B buyers. Worth tracking if you sell to professionals.
- Microsoft Copilot. Built on Bing. Matters for businesses selling into corporate environments where Edge and Microsoft 365 are the default.
If you are a Melbourne cafe or tradie, the first entry is nearly the whole game. If you sell software or professional services nationally, the next three carry real weight.
What gets a page cited: the patterns we see
We cannot read the engines’ retrieval code, and anyone who claims a formula is guessing. What follows is pattern, not proof, drawn from the client pages that do get cited and the ones that do not.
- The answer is stated first. A page that opens with a direct two-sentence answer to the question in its heading, then supports it, gets lifted far more often than a page that builds to the point.
- Chunks are self-contained. Each heading introduces a passage that makes sense on its own. Engines extract passages, not pages.
- The page contains something original. A price range from your own quotes, a figure from your own data, a photo of your own work. Engines have no reason to cite the fourth rewrite of the same generic advice.
- Entity facts match everywhere. Business name, address, phone, founding year and service list agree on your site, your Google Business Profile, directories and social profiles. Inconsistency reads as low confidence.
- Authorship is visible. A named person with a real role and a short bio, and a date that is honest.
- Structured data describes what is on the page. Organisation, LocalBusiness, Service, FAQ and Article markup that mirrors the visible content, not markup that invents it.
- The page already ranks. Pages on the first two pages of Google are cited far more often than pages that are not. GEO is not a shortcut around ranking.
A useful test: paste your page’s opening paragraph into any assistant and ask it to answer the page’s title question from that text alone. If it cannot, a retrieval engine will not either.
GEO for local Melbourne searches
Local answers behave differently. When someone asks an assistant for a dentist in Hawthorn or an electrician who works weekends in the west, the engine leans heavily on Google Business Profile data, reviews and the suburb pages it can find. The local pack has not gone anywhere; the AI layer sits beside it and draws from the same sources.
- Keep the Business Profile complete: every service listed, hours accurate, categories right, photos recent, questions answered.
- Earn reviews that mention the service and the suburb in the customer’s own words. Engines quote review language when summarising a business.
- Give each suburb you genuinely serve its own page with real detail: parking, response times, the jobs you did there. A city name dropped into a template does not get cited.
- Make sure name, address and phone are identical across the site, the profile and the major directories.
Our local SEO and Business Profile work covers this ground; it is the same foundation that AI answers draw from.
How to measure GEO
Measurement is the part most agencies skip, because none of the engines publish a clean report. You can still build a reliable picture from four sources.
- GA4 referral traffic. Visits from chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com arrive as referrals. Build a channel group or an exploration that isolates them, and watch sessions and conversions month on month.
- Search Console. Google folds AI Overview impressions and clicks into the normal Search performance report; it does not separate them. Treat a rise in impressions with falling clicks on informational queries as a sign that an Overview is answering above you, and check the query live.
- Prompt tracking. Keep a list of twenty or thirty prompts your customers would actually ask, run them monthly in each engine, and record whether you are cited and in what position. It is manual, and it is the only direct citation measure available.
- Server logs. Crawlers such as GPTBot, ClaudeBot, PerplexityBot and Google-Extended identify themselves in your logs. If they never fetch a page, that page cannot be cited from a live retrieval.
Report the four together with leads and revenue. A citation that sends no one is a vanity metric; the aim is assisted conversions you can attribute.
A 30-day GEO starter plan
Thirty days is enough to set the foundation if the technical basics are already sound.
- Week 1. Fix entity consistency across the site, the Business Profile and the top directories. Confirm robots.txt is not blocking the AI crawlers you want. Set up the GA4 referral view.
- Week 2. Pick your ten most important pages. Rewrite each opening so the answer comes first, and add one piece of original evidence per page.
- Week 3. Add or correct structured data so it mirrors the visible content. Add visible authorship and honest dates. Write the prompt list and take the first citation baseline.
- Week 4. Publish one page that answers a question your customers ask and nobody in Melbourne has answered properly, with your own data in it. Then re-run the prompts.
If you would rather have this done, our AI SEO service runs exactly this programme, and the AI SEO audit checklist is the diagnostic we start with.