We asked ChatGPT about our company and a different company came back
Whether the answer engine does not know us, or knows us but has no document to use as evidence, are completely different causes. The audit starts by separating those two.
AEO AUDIT
AEO audit —
what is AI saying
about you right now
You get a diagnosis before you get a proposal. We put real questions to the major answer engines and record whether the brand is cited, in what sentences it is described, and who currently holds that slot.
In one paragraph
An AEO audit measures how a brand is currently handled by AI answer engines, by putting real questions to them. For each question we record separately whether the brand was mentioned, whether it was cited as a source, whether the description matches the facts, and which competitor holds that slot. It is a procedure for starting from observation rather than assumption, and improvement work is decided only after this record exists.
A proposal written without an audit generally ends up generic. When nobody knows what the problem is, the only available conclusion is to write more content. Navirang records a baseline first and compares every subsequent change against it.
Takes 1-2 weeksFree audit availableFive report deliverables
Last verified
WHEN YOU NEED THIS
When you have these problems,
this is the work you need
These are the situations we hear repeatedly in consultations. If any of them apply, start by measuring.
Our search rankings are good but we never appear in AI answers
Search ranking and AI citation are not the same pipeline. It is common for a page that ranks well to be structured in a way that cannot be cut into a citable unit.
AI gets our pricing, services or company history wrong
You cannot correct a wrong answer without finding which source it came from. The audit records the incorrect sentence together with the source it appears to have drawn on.
We have no way to verify the numbers in an agency proposal
A citation rate becomes an entirely different value depending on the denominator. Fixing the measurement conditions in writing first means any later figure can be compared on the same basis.
WHAT WE DO
What Navirang
actually does
Written as units of work rather than abstract proposals. The scope of an engagement is set from this list.
Question set design
We settle on 20-40 real questions the brand should be the answer to, composed across industry, purchase stage and question type, with the wording unchanged for the duration of the measurement.
- Separate informational, vendor-search, comparison and conditional questions
- Separate brand-named questions from unbranded ones
- Fix the question wording for the measurement period
Live measurement in the answer engines
We put the designed questions directly to the answer engines and collect the responses. Because logged-in state and personalization move the results, we work in logged-out sessions and repeat on the same day and time of day.
- ChatGPT, Google AI Overviews, Perplexity, Claude, Naver AI Search, Copilot, Gemini
- Repeated measurement per question to rule out one-off responses
- Response text and source lists preserved
Mentions and citations recorded separately
Being named (a mention) and being linked as a source (a citation) are different outcomes. Mixing them leaves you unable to tell what to improve, so we count the items apart.
- Mentioned or not / cited or not
- Position within the source list
- Factual accuracy of the description
- Competitors named alongside
Technical accessibility checks
We confirm the site is in a state where answer engines can read it. What looked like a content problem has frequently turned out to be a crawler block or a rendering method.
- AI crawler policy in robots.txt
- Whether body text exists in the HTML source (JavaScript dependence)
- Presence of structured data and whether it matches the visible page
- Sitemap, canonical URLs and index status
Mapping the competitive picture
We compile the domains repeatedly cited for the same questions. It is common for the slot to be held by a blog, a community or a news outlet rather than a competitor site, and the response differs in that case.
- Cited domains tallied per question
- The visibility gap against your own site
- Slots held by third-party media, marked separately
Improvement priorities
We order the problems found by effect against cost. Items you can handle internally are marked as such and kept out of the engagement scope.
- Split into immediate, short-term and medium-term
- Items you can handle internally marked
- Metrics to confirm at the next measurement
PROCESS
In what order
does it run
What you receive at each stage is stated alongside it. Durations are the working time Navirang controls; they are not a promise about when results appear.
- 01 1-2 days
Scope
We confirm the site URL, the main services, and the brands you regard as competitors. No documents are needed up to this point.
Audit scope definition
- 02 2-3 days
Question set design
We finalise 20-40 questions the brand should be the answer to. The finalised questions are reused as-is in later re-measurements.
Question set document
- 03 3-5 days
Measurement
We put the questions to the answer engines, collect responses, and record mentions, citations, accuracy and competitors item by item.
Observation record
- 04 1-2 days
Technical checks
We check crawler accessibility, rendering, structured data and index status to separate content problems from technical ones.
Technical checklist
- 05 1-2 days
Report
We write up the observations, our reading of the causes and the improvement priorities, and walk you through them.
AEO audit report
DELIVERABLES
What you
receive
We do not do work that ends in conversation. The documents below remain, and become the baseline for the next measurement.
Question set document
Every question used in the measurement and how they were classified. It becomes the basis for re-measurement, so the client keeps it.
Observation record
A table recording mention, citation, position, accuracy and competitors for each question and engine combination. You get the raw observations, not just summary figures.
Technical checklist
Crawler policy, rendering, structured data and index status, each marked pass or fail.
Misinformation list
Sentences answered contrary to fact, alongside the source they appear to have drawn on.
Improvement priorities
An execution list ordered by effect against cost. Items you can handle yourself are marked.
We fix the measurement conditions first
- Answer engines
- ChatGPT, Google AI Overviews, Perplexity, Claude, Naver AI Search, Copilot, Gemini
- Number of questions
- 20-40, split by industry, purchase stage and question type
- Session conditions
- Logged-out sessions, fixed question wording, repeated on the same day and time of day
- Recorded items
- Mentioned / cited / position in source list / accuracy of description / competitors named alongside
- Duration
- 1-2 weeks
Navirang applies the same method to its own site, measuring 40 questions x 7 answer engines = 280 cells as a baseline. That record is published with its conditions in our eight-week improvement log.
HOW IT CONNECTS
How it connects
to the other work
Our work moves as one piece. SEO builds the foundation for being found by search engines, AEO raises the odds of that information being cited in an answer, structured data helps machines understand the facts, and content supplies the evidence there is to cite.
| Area | Relationship to this work |
|---|---|
| Citable content design | Filling the missing document the audit found you had nothing to answer with |
| Structured data | Filling the missing machine-readable facts the audit found |
| Citation monitoring | Re-measuring the baseline the audit created, under the same conditions |
| SEO | Fixing the foundation when the audit surfaces indexing or crawl problems |
| AI-ready websites | Rebuilding the site structure when the rendering method itself is the cause |
| AEO by industry | When the question set is built from that industry question types |
FAQ
Frequently asked questions
Q What does an AEO audit measure?
A Whether an answer engine mentions the brand when answering a given question, whether it cites it as a source, whether the description matches the facts, and which competitors appear in the same slot. Search ranking is not measured, because ranking and citation are different metrics.
Q How long does the audit take?
A One to two weeks. Most of the time goes into question set design and the measurement itself, and it varies with the number of questions and engines. We reply to enquiries within one business day.
Q What is the difference between the free and the paid audit?
A The free audit covers a core question group, checking current citation status and technical accessibility and delivering a report. Extending the question set across an industry, or tracking competitor share on a schedule, is a separate engagement. You keep the free audit report either way.
Q Can the audit guarantee ranking or visibility?
A No. Answer engine responses differ for the same question depending on the moment and the session, and each engine selection criteria are not public. What Navirang promises is that change can be confirmed by repeated measurement under the same conditions. OpenAI likewise states in its official documentation that there is no way to guarantee top placement.
Q Do we need to prepare materials in advance?
A No. The site URL and the names of the brands you regard as competitors are enough to start. Auditing from public information only is in fact more accurate, because the answer engines see the site under the same conditions.
Q Do we have to sign with Navirang after the audit?
A No. The report marks the items you can handle internally. Not tying work you do not need to hand over into a contract is a principle here.
What is AI saying about your brand right now?
Send us a URL and the free audit finds out. We agree the question set and measurement conditions first, and send the report with the raw observations attached.
We reply within one business day.