KoreLens

KoreLens playbook

New site? How to get found by AI.

If the AI answers about your business describe a different company — or none at all — the engines simply don’t have your site yet. New sites start invisible to AI: models are trained on a snapshot of the web from before you existed, and search-backed answers can only cite pages already indexed. Neither is a verdict on your business, and both respond to the same free, unglamorous steps below — the ones we use ourselves.

Why AI answers about you can describe someone else

When an AI is asked about a name it has never seen, it doesn’t say nothing — it reaches for the nearest name it can corroborate, which may be a similarly-spelled company on the other side of the world. The cure is not arguing with the model. It is giving every engine something real to index and corroborate, so the nearest corroborated name becomes you.

Step 1

Make sure AI crawlers can read your site at all

What this reaches: Everything downstream. A crawler that is blocked, challenged, or walled out can never index you — this step gates all the others.

  1. Remove password walls, "coming soon" gates, and platform preview modes from the pages you want read.
  2. Check robots.txt does not disallow the retrieval crawlers — OAI-SearchBot (ChatGPT search), Claude-SearchBot, PerplexityBot, Googlebot, bingbot — or the user-requested agents that fetch a page when a person asks (ChatGPT-User, Claude-User, Perplexity-User).
  3. Firewall and CDN bot challenges can block crawlers your robots.txt allows — if you use one, check its verified-bot settings too.

How you know it worked: The free check reads your live robots.txt and names, bot by bot, who can and cannot read you.

Clock: Immediate — the moment the block is gone.

In KoreLens: Crawler access is scored in every free check. Change robots.txt, re-run the check, watch the section flip.

Step 2

Register with Google Search Console and submit your sitemap

What this reaches: Google’s index — the foundation Google Search, AI Overviews, and AI-mode answers build on, and the grounding source for Gemini answers that search the web.

  1. Verify your domain at search.google.com/search-console (a DNS record where your domain is hosted, or an HTML file upload).
  2. Submit your sitemap (usually /sitemap.xml) under Indexing → Sitemaps.
  3. Use URL Inspection on your homepage and key pages, then Request indexing.

How you know it worked: URL Inspection reports “URL is on Google”, and a site:yourdomain.com search starts returning your pages.

Clock: Verification is instant; first indexing usually takes days.

In KoreLens: Once you have data, export the Performance CSV and import it into KoreLens — index status sits next to what the AIs actually say.

Step 3

Register with Bing Webmaster Tools — one click if you did Search Console first

What this reaches: Bing’s index — what Microsoft Copilot answers from, and a crawl source smaller assistants reuse (DuckDuckGo’s results are Bing-backed). For a new site this is the most commonly skipped step.

  1. Sign in at bing.com/webmasters and choose “Import from Google Search Console” — verification and sitemaps carry over in one step.
  2. No Search Console yet? Verify by DNS or XML file and submit /sitemap.xml directly.

How you know it worked: A site:yourdomain.com search on Bing returns your pages, and Bing Webmaster’s reports show pages indexed.

Clock: Days, typically.

In KoreLens: Bing Webmaster’s exported report can be imported into KoreLens and kept alongside your AI answers, clearly labelled as entered by you.

Step 4

Invite a re-crawl with IndexNow when pages change

What this reaches: Bing and the other IndexNow participants re-crawl changed pages sooner. Google does not participate — this never replaces Search Console.

  1. Host a key file at https://yourdomain.com/<key>.txt whose body is the key itself.
  2. Submit changed URLs to api.indexnow.org — or let KoreLens do it from your dashboard: it verifies your key file with a real fetch first, and reports back the exact HTTP status it received.

How you know it worked: A 200/202 response means the submission was received — an invitation to re-crawl, never proof of indexing, and KoreLens never records it as more than that.

Clock: Submission is immediate; the re-crawl happens at each engine’s discretion.

In KoreLens: The dashboard’s IndexNow form does the key verification and submission for you, and logs the real response.

Step 5

Publish one page that says plainly what you are

What this reaches: Any engine answering “what is <your name>” needs one crawlable paragraph to quote. When it has none, it reaches for the nearest name it can corroborate — which may be a similarly-spelled business that isn’t you.

  1. Write a simple About page: your name, what you sell, who it’s for, where you’re based, and since when — in prose a stranger could quote whole.
  2. Link it from your footer and include it in your sitemap, so every crawler that finds you finds it.
  3. If your name is a near-miss of another company’s, this page does the disambiguating — make what you are unmistakable rather than arguing about what you are not.

How you know it worked: Once the page is indexed, re-run the free check and read what the engines quote.

Clock: An hour to write; it starts working once indexed.

Step 6

Make other sites agree on who you are

What this reaches: Engines weight corroboration: a claim only your own site makes is weaker than the same claim three independently-known sites repeat.

  1. Create only the profiles you will actually maintain — a LinkedIn company page, an X profile, your platform’s app or partner directory — each with the identical one-line description and a link to your domain.
  2. Then list those profiles as sameAs in your site’s Organization schema. Only profiles that really exist: an empty sameAs is honest, a padded one is not.

How you know it worked: The profiles appear in ordinary search within days, and the free check reads the schema on your site on every run.

Clock: An afternoon to set up; corroboration compounds from there.

Step 7

Never write your name without context

What this reaches: Retrieval snippets travel without your homepage around them — the words immediately next to your name are often all an engine quotes.

  1. Everywhere your name appears — page titles, social bios, directory entries — pair it with what you do: “<your name> — <what you do, for whom>”.
  2. Keep the pairing identical everywhere. Consistency is what turns a spelling into an entity.

How you know it worked: Search your own name in a private browser window and read the snippets, not the pages.

Clock: Ongoing habit, zero cost.

Two clocks — be honest with yourself about both

  • Search-backed answers: Answers marked “searched the web” can change within days to weeks of your pages being indexed. This is the clock every step above moves.
  • Built-in model knowledge: A model’s built-in knowledge — answers given without searching — changes only when the model is retrained: months, on the provider’s schedule. Nothing you or any tool does moves that clock, and anyone who says otherwise is selling you the weather.

What none of this guarantees

None of this makes an AI mention or recommend you, and nobody can sell you that. These steps make your site readable and your identity corroborable — readiness, never outcome. The free check measures the part you control; the 7-day re-check shows what actually changed.

Start where the playbook starts

The free check reads your live robots.txt, your schema, and what the AI engines actually say about you today — the before picture every step above moves.

Run the free check — about a minute, no signup