21 checks before
you start.
A practical self-audit for Singapore and Malaysian businesses. Run through each check, record what you find, then decide what to fix first. Download the blank checklist as a CSV to track your progress.
Short answer first.
Work through the four groups and record confirmed findings, items needing review, not checked or not applicable. Add the evidence and a person responsible for the next action. Some checks will not apply to your business; this is a practical worksheet, not a pass mark for AI inclusion.
This is a self-check, not a live test or ranking forecast. Passing every check does not guarantee that an AI platform will name or cite your business. Failing a check identifies a gap worth investigating.
If you want a professional audit with dated AI answer evidence across four platforms, see the Singapore audit or the US audit.
Buyer question coverage.
Does your website answer the questions buyers ask before choosing a provider?
On a small screen, swipe each table sideways to read all four columns. Keyboard users can focus a table and use the arrow keys. The CSV provides an editable view.
| Check | What to inspect | Acceptable evidence | Next action if missing |
|---|---|---|---|
| Service page answers the buyer question | Does a specific page answer the question a buyer would ask before contacting you? | The page names the service, location and constraint in the heading and first paragraph | Write or rewrite the service page to answer the question directly |
| One page per buying decision | Is there a clear lead page for each important buying decision, with supporting pages linked to it? | A service-to-page map and relevant internal links; inspect actual query/page data before diagnosing competition | Clarify the lead page; consolidate only genuinely overlapping intent |
| Buyer language matches page content | Does the page use the words and phrasing buyers actually use? | Heading and body text match real enquiry phrasing from WhatsApp, calls or Search Console queries | Rewrite headings and body using real customer language |
| Service scope is accurate | Does the page describe only services you actually provide, in locations you actually serve? | Every service and location named is verifiable by the business | Remove or correct unsupported claims |
| Clear next step for the buyer | Can a visitor find and use the relevant enquiry, booking or quote route? | A usable contact route tested on phone and desktop, with the right destination | Repair or make the relevant contact route easier to find |
Public evidence and entity accuracy.
Can search engines and AI platforms verify the facts about your business?
| Check | What to inspect | Acceptable evidence | Next action if missing |
|---|---|---|---|
| Google Business Profile is accurate | Are business name, address, phone, hours and categories correct? | Profile facts match the business website and are currently verifiable | Correct wrong facts via Google Business Profile manager |
| Consistent name, address, phone across directories | Do the public sources relevant to your business show verified current contact details? | Checks of the actual profiles and publishers customers use; record each source and date | Request corrections for verified mismatches; there is no minimum directory count |
| Structured data identifies the business | If structured data is present, does it accurately describe the visible business information? | Valid markup consistent with visible facts; no special AI schema is required for Google AI features | Correct inaccurate markup; add useful descriptive markup only where appropriate |
| Google reviews are accurate and current | If your website repeats a Google rating or review count, does that claim match the checked profile? | The source profile, capture date and any corresponding website claim are recorded | Correct stale claims; reviews are not an AI eligibility requirement; never buy or fabricate them |
| No wrong facts in existing AI answers | Have you checked whether AI answers currently say anything incorrect about your business? | Run your main buyer questions on ChatGPT, Google AI, Gemini and Perplexity; record any wrong facts with dates | Correct the source where the wrong fact originates; see the wrong information guide |
Crawl and discovery.
Can search engines and AI search crawlers actually access your pages?
| Check | What to inspect | Acceptable evidence | Next action if missing |
|---|---|---|---|
| Pages are crawlable | Can the intended search crawlers reach public service pages through both robots rules and hosting controls? | Review the policy and actual accessible content; use authenticated crawler evidence where available | Investigate unintended blocking with the host while preserving deliberate security and privacy controls |
| Pages are indexed | Are your service pages in the Google index? | Google Search Console URL Inspection shows the page as indexed | Investigate indexing issues; request recrawl after fixing |
| No accidental noindex | Are pages intended for public search unintentionally marked noindex in HTML or HTTP headers? | Check both the page directive and response headers against the intended publication policy | Remove only an unintended exclusion after review, then verify the public response |
| AI search crawlers are not blocked | Do intended search crawlers have access under robots rules and hosting policy? | Review robots rules and validate actual bot requests using published IP ranges or verified-crawler metadata | Investigate unintended blocks; keep search, training and user-request controls separate |
| Content is not behind a login wall | Can intended visitors and search systems read the public service facts without signing in? | Important facts are available in accessible public page content; verify the rendered or retrieved content | Publish intended public facts accessibly while keeping private information protected |
| XML sitemap is submitted and current | Does the sitemap accurately list current canonical pages intended for search? | A valid accessible sitemap, with submission status recorded if submitted | Correct stale entries and optionally submit; submission aids discovery but does not guarantee indexing |
| Internal links connect service pages | Do your service pages link to each other and supporting pages? | Service pages include contextual links to related pages | Add relevant internal links between service and supporting pages |
Measurement and conversion.
Can you tell whether AI visibility produces real enquiries?
| Check | What to inspect | Acceptable evidence | Next action if missing |
|---|---|---|---|
| Enquiry source tracking is active | Can contact events and confirmed enquiries be traced to an actual page or channel? | Working destination tests, available analytics source/page data and separately confirmed contact records | Test the enquiry route and capture usable source context; a click or URL parameter alone is not an enquiry |
| Search Console is connected and reporting | Is Google Search Console verified and showing data? | Search Console shows clicks, impressions and average position for the last 28 days | Verify the property in Search Console |
| AI answer checks are dated and repeatable | Are you running the same buyer questions on the same platforms with dates recorded? | A fixed panel of questions, checked periodically, with date, platform and status saved | Establish a locked question panel and check schedule. See how to choose questions |
| Named and cited are recorded separately | Do your records distinguish between being named and having your website cited? | Each check records named and cited as separate fields. See how we measure | Update your tracking to separate named from cited |
What the checks
tell you.
The distinction matters: crawlable means search engines can access the page. Indexed means Google has included it. Named means an AI answer identified your business. Cited means the answer linked your own website as a source. Enquiry means a real person contacted you. Each is a separate observation.
For Google AI features in Search, an eligible supporting page must be indexed and eligible for a search snippet. No special AI schema, text file or additional optimisation is required. See Google's AI features documentation.
Other AI platforms (ChatGPT, Perplexity) have their own retrieval systems. OpenAI documents separate search (OAI-SearchBot), training (GPTBot) and user-directed (ChatGPT-User) crawlers with independent controls. See OpenAI crawler documentation. Allowing training crawlers is not a prerequisite for search visibility. Do not weaken website security to accommodate training bots.
This checklist helps you identify gaps worth investigating. It is not a ranking score, a guarantee of AI inclusion, or a substitute for a professional audit with dated evidence across four platforms.
Track it yourself.
The CSV has the same rows with blank Status, Evidence and Owner columns. Open it in any spreadsheet application and fill in your findings.
What's holding
your business back?
A website that needs work. Customers who can't find you. A team running on spreadsheets and messages. Start there.