Answer Engine Optimisation
Become the source AI answers cite.
We help brands become more visible and cited in ChatGPT, Google AI Overviews, Gemini, and Perplexity.
When someone asks an AI assistant for a recommendation, the assistant answers in prose and cites a handful of sources. Answer Engine Optimisation (AEO) is the work of becoming one of those cited sources. It overlaps with SEO but is not the same job: search engines rank pages, answer engines extract and attribute claims.
What you receive
Buyer-question research
The actual questions your buyers type into assistants, gathered from sales conversations, support tickets, and search data — not invented keyword lists.
AI visibility audit
A run of those questions across ChatGPT, Google AI Overviews, Gemini, and Perplexity, recording which brands are named and which sources are cited.
Competitor analysis
Which competitors are being cited, on which questions, and which specific pages the assistants are pulling from.
Technical checks
Whether AI crawlers can actually reach and read your site: robots rules, bot firewall rules, server-rendered content, structured data, and canonical URLs.
Content plan
A prioritised list of pages to write or rewrite, mapped to the questions where you are currently absent.
Answer-focused content
Pages written so a single paragraph can be lifted and attributed — direct answers, clear headings, and claims that stand on their own.
Citation tracking
A recurring record of where you appear, so movement is measured against a baseline rather than guessed at.
How we work
Discover buyer questions
We collect the real questions your buyers ask, in their words, and turn them into a fixed prompt set we can re-run over time.
Measure current visibility
We run that prompt set across each assistant and record the brands named and the sources cited. This becomes your baseline.
Find technical and content gaps
We check whether AI crawlers can read your site at all, then identify the questions where you have no page worth citing.
Implement improvements
We fix the technical blockers and write the answer-focused pages, or hand your team a specification precise enough to build from.
Track visibility and citations
We re-run the same prompt set on a schedule and report the change against baseline.
Sample Audit
A real snapshot, not a mock-up.
Five buyer questions from the Malaysian software market, run in August 2026. For each we recorded the brands named in the answer and the domains cited as sources. This is the retrieval layer that Google AI Overviews and ChatGPT search draw from; a full engagement additionally runs the same prompts inside each assistant and records those answers.
| Buyer question | Brands named | Domains cited |
|---|---|---|
| Best POS system for retail business in MalaysiaSiteGiant ranks first with its own blog. StoreHub is named, but only inside other companies' listicles — none of the cited sources is storehub.com. |
|
|
| Best payroll software for small business MalaysiaPayrollPanda is named first and described most favourably — but every cited source is a competitor's blog. The recommendation is being made about them, on someone else's page. |
|
|
| Best HR software for SME Malaysia employee benefitsHealthMetrics does not appear at all, despite operating in exactly this category. Competitors who publish comparison content own the answer. |
|
|
| Best solar panel installer for home Malaysia 2026Verdant Solar is absent. Trexon captures the answer twice with its own comparison pages — a smaller player outranking larger installers purely on published content. |
|
|
| Best ecommerce platform for online store MalaysiaSiteGiant is cited twice from its own domain. This is what winning looks like: the brand is both the answer and the source. |
|
|
Three improvements this snapshot points to
Publish the comparison page yourself
In four of the five questions, the cited source is a comparison or "best of" article. Where that article belongs to a competitor, they control how you are described. The single highest-leverage fix is publishing an honest comparison page on your own domain covering the category you compete in.
Answer the question in the buyer's words
The pages being cited use the buyer's phrasing in the heading and answer within the first paragraph. Product pages written in feature language are rarely cited, because there is no self-contained sentence for an assistant to lift and attribute.
Being named is not being cited
PayrollPanda is named in the payroll answer but no cited source is their domain. Being mentioned inside a competitor's article means the click, the attribution, and the framing all go elsewhere. Closing that gap means owning at least one cited source per priority question.
How results are measured
Every engagement starts with a baseline: a fixed set of buyer prompts, run across each platform on a recorded date, with the brands named and sources cited written down. Progress is measured by re-running that identical prompt set on a schedule and comparing against the baseline.
Prompt coverage
The share of your prompt set where your brand is named at all.
Citation rate
How often your own domain is the cited source, rather than a directory or a competitor.
Share of voice
How often you appear relative to the named competitors on the same prompts.
Referral traffic
Sessions arriving from assistant referrers, tracked separately from organic search.
What we do not promise
Citations and rankings cannot be guaranteed. No agency controls how ChatGPT, Gemini, Perplexity, or Google AI Overviews select and attribute sources — these systems change without notice and answers vary between users and sessions. What we guarantee is the work and the measurement: a documented baseline, the technical and content changes, and honest reporting of what moved and what did not.
Start with a pilot
A four-week paid pilot on a single product line or service area: 20 buyer prompts, a full baseline across four platforms, a technical audit, and three prioritised fixes implemented. You keep the prompt set and the baseline data whether or not you continue.
- 20 buyer prompts, agreed with you before we start
- Baseline measured across ChatGPT, Google AI Overviews, Gemini, and Perplexity
- Technical crawlability audit with specific fixes
- Three prioritised improvements implemented
- A re-measure at week four against the same prompts
Questions about AEO
What is the difference between AEO and SEO?
- SEO optimises for ranking a page in a list of blue links. AEO optimises for being quoted and attributed inside a generated answer. The work overlaps — both need a crawlable, well-structured site — but AEO puts more weight on writing self-contained claims an assistant can lift, and on being corroborated by third-party sources the assistant already trusts. A page can rank first on Google and still never be cited by ChatGPT.
Which platforms do you cover?
- We measure ChatGPT (search mode), Google AI Overviews, Gemini, and Perplexity as standard. These four use different indexes — ChatGPT leans on its own crawler and Bing, Google AI Overviews uses Google's index, and Perplexity runs its own — so visibility on one does not imply visibility on another. Claude is available on request.
How do you measure results?
- We fix a prompt set at the start of the engagement and record a dated baseline of which brands are named and which sources are cited on each platform. Progress is the change in that same prompt set over time, reported as prompt coverage, citation rate, and share of voice against named competitors. Because assistant answers vary between sessions, we run each prompt multiple times and report the pattern rather than a single result.
How long does it take to see results?
- Technical fixes — making your site readable to AI crawlers — can change what assistants retrieve within days to a few weeks. Content and authority work is slower, typically three to six months, because assistants weight corroboration from sources beyond your own site. Any agency promising citations in weeks is describing something they do not control.
Can you guarantee citations?
- No. No one can. The companies running these assistants change retrieval and attribution without notice, and identical prompts can produce different answers in different sessions. We guarantee the work and the measurement — a documented baseline, the implemented changes, and honest reporting — not a specific placement in a system we do not control.
Request a free 5-prompt visibility snapshot
Send us five questions your buyers actually ask. We run them across the major assistants and send back what we find — which brands are named, which sources are cited, and where you stand. No cost, no commitment.
Request your snapshot