Why nowThe auditThe reportLive casesThe fixPricingFAQ
Free snapshot
StartupRaise PapersADMIT ONE Every service, one thread VISIT THE DESK →
The buyer journey moved into the answer box

Your buyers ask AI.It names your rivals.Let's fix that.

ChatGPT, Perplexity, Gemini and Claude answer thousands of buying questions every day. The Audit shows you exactly what they say about you, which sources they trust, and how to get in. Built with AI. Checked by a human. No fake accounts.

Get your free snapshot Full audit ₹1,999
Every answer screenshotted Delivered over WhatsApp Flat fee, no retainer trap
What your buyer sees
ChatGPTPerplexityGeminiClaude
>
You weren't mentioned. Anywhere.Missing
answer built from 6 sources · none of them include you
0
AI models tested: ChatGPT, Perplexity, Gemini, Claude
0
real buyer prompts, built around your category
0+
answers captured as screenshots, you against rivals
0
ruthless report with a 30-day plan, in days not weeks
Why now

How an AI answer gets made.

Three steps, none of them magic. Once you see the machine, you can win it.

01

A buyer asks

Best X in India. What should I buy. Is this brand any good. Thousands of these questions a day, and the asker trusts whatever comes back.

02

The model goes looking

It runs a live search and reads a handful of sources it trusts: listicles, comparison pages, communities, directories, and your own website.

03

The answer names brands

Whoever is in those sources gets recommended. Whoever isn't doesn't exist. There is no page two to scroll to.

Communities Listicles Directories Reviews News Your site - - -> The answer
Not in the sources, not in the answer. That's the whole game.
The audit

Thirty questions. Four models.
Zero guessing.

Every claim in the report is a screenshot you can open.

Four models, thirty real prompts

The questions your buyers actually ask, run on ChatGPT, Perplexity, Gemini and Claude. Where you're mentioned, where you rank, what each one believes about you.

ChatGPT42%
Perplexity27%
Gemini8%
Claude35%
mention rate across the 30-prompt set · sample data

Screenshots, not summaries

Every answer captured exactly as your buyer sees it. No interpretation in between.

prompt 04 · chatgpt · best in category · 2 rivals named
prompt 12 · perplexity · is it legit · thin answer
prompt 18 · gemini · best for gifting · you absent
prompt 23 · claude · alternatives to · you absent
every claim in the report opens its screenshot
120+

answers captured per audit, every single one saved as a screenshot you can open

The citation map

The domains behind each answer, and the exact gaps where rivals appear and you don't.

listicle, category rounduprivals ✓ · you ✗
community threadrivals ✓ · you ✗
startup directoryrivals ✓ · you ✗
your own websiteunreadable to models

The 30-day plan

Ranked by effort against impact, written so you could run it without me. Most founders hand it straight back and let me run it instead.

WK 1

Machine-readable

Schema, llms.txt and plain definitions live on your site
WK 2

Proof profiles

The directories and registries the models treat as verification
WK 3

Source pitches

Real angles to the exact pages the models cited
WK 4

Earn citations

Comparison content on your site, then the full re-test
The deliverable

The report, exactly as it lands on WhatsApp.

A visibility score, the share of every answer, the sources behind it, and the proof. Sample layout shown with placeholder data.

ai-visibility-report · your-brand · june 2026
Visibility score
0out of 100
Mentioned in 10 of 30 buyer prompts across four models
Score over time
re-tested monthly · movement publishes to the live cases
Share of answer · your category
Rival A64%
Rival B51%
Rival C38%
You11%
Where the answers come from
category listicleyou're absent
community threadyou're absent
startup directorylisted, thin
What each model believes
chatgptknows you exist
perplexityconfuses your name
gemininever heard of you
claudedescribes you, vaguely
Fix list · first 30 days
schema + llms.txtwk 1
profiles + directorieswk 2
source pitcheswk 3-4
p.04p.09p.13p.18p.27p.31
Every number above links to a screenshot inside the report. No adjectives, just evidence.
Watched in public

Two live audits. Real brands. No script.

No invented numbers and no borrowed logos. Case 01 is this very website. Case 02 is a real pre-launch D2C brand, name withheld until launch. Both update here, monthly, whichever way the answers move.

LiveCASE 01 · STARTED 10 JUN 2026ServicesSelf-audit
This website
Day-zero baseline · measured, not estimated
✗The category answer is owned by incumbent listicles and service sites. This brand: absent from all of them.
✗Brand searches drown under bigger names that share its keywords.
✗Machine-readable footprint on day zero: none.
Schema, llms.txt and a plain-text definition shippedDone · 10 Jun
Company profile, Bing index, founder directoriesIn progress
Honest comparison post aimed at the category queryDue · wk 3
Pitches to the exact pages the models citedDue · wk 3-4
Full 30-prompt re-test, published on this pageDue · Jul 2026
The score and the movement publish here after the July re-test, up or down. That's the deal.
LiveCASE 02 · STARTED 10 JUN 2026D2C · pre-launchName withheld
Premium silver jewellery brand
Day-zero baseline · measured, not estimated
✗The category answer is built almost entirely from listicles the rival brands wrote about themselves, each ranking itself first.
✗One rival appears in nearly every list the models read.
✗The brand's specialist material is barely contested in India: an open query waiting to be owned.
30-prompt baseline across all four modelsIn progress
Definition layer and schema live before launchDue · wk 2
Proof profiles: directories, registries, company pageDue · wk 2-3
Material-education page to own the category termDue · wk 3-4
Re-test with before and after screenshots, published hereDue · Aug 2026
Brand name and screenshots publish at launch. Until then, the steps and statuses above are the case study.
The fix

How I get you in.

Five levers, all of them legitimate. The audit tells us which to pull first.

01

Your site, made machine-readable

Structured data, llms.txt, and a plain-language definition of what you do that a model can quote. Most startup sites are written for humans only, so the models skip them.

02

Proof-of-existence profiles

The directories and registries models treat as verification that you're real. Built properly once, maintained, and consistent with your site word for word.

03

The exact sources models cite

The audit shows which listicles and comparison pages built the answer. Each one gets a real pitch with a real angle. If a page ranks your category, you belong on it.

04

Community answers, under your real name

Genuinely useful replies where your buyers already ask, signed by you. Models weight these heavily, and so do the humans reading them.

05

Content that earns citations

One honest comparison page on your own site can become the source the models quote for your whole category. That's the compounding asset.

The line that doesn't move

No fake accounts. No bought personas. No undisclosed posting. Tactics like that get brands banned from the platforms and burned in investor diligence. Everything here survives daylight.

How it works

Four steps. One thread.

01

Message me on WhatsApp

Send your website and your three closest rivals. Fifteen minutes of questions about your buyer, in writing.

02

Confirm and pay

I confirm scope and the flat fee, then send a payment link right in the same thread.

03

I test and build

The AI runs the prompts at scale. I check every screenshot myself and write the plan by hand.

04

It lands in the thread

The full report in days, not weeks. We walk through it together and decide what to fix first.

Pricing

Flat fees, shown upfront.

Start with the audit. The retainer exists for founders who'd rather I run the plan.

Launch month only. Lock the founder rate before the price returns to the struck figure.
Most founders start here One-off · delivered in days
AI Visibility Audit
Where you stand when AI answers your buyers, and exactly how to change it.
  • 30 real buyer prompts across ChatGPT, Perplexity, Gemini and Claude
  • A screenshot of every answer, you against three rivals
  • The citation source map, and where you're missing
  • The 30-day plan, ranked by effort against impact
  • Walked through together on WhatsApp
Price this month
₹9,999₹1,999
Begin
After the audit · monthly
Visibility Retainer
I run the plan, you watch the answers change. Month to month, cancel anytime.
  • The full 30-prompt re-test every month, screenshots included
  • Source placements pitched and executed, the honest way
  • Profiles, directories and site fixes built and maintained
  • A before and after report you can forward to anyone
Launch rate, locked while you stay
₹19,999/ month
Talk first

Find out what AI says about you.

Three prompts, real screenshots, one WhatsApp message. Free, within a day, no call needed.

Get the free snapshot
No payment. No call. Just send your website.
Questions

Straight answers.

What exactly do I get in the audit? +
A report covering 30 real buyer prompts tested across ChatGPT, Perplexity, Gemini and Claude, a screenshot of every answer, your mentions against three rivals, the domains each model cited, the gaps where you're missing, and a 30-day plan ranked by effort against impact.
How is this different from SEO? +
SEO earns you a position in a list of links. AI answers skip the list and name a few brands directly, built from the sources the model retrieves and trusts. The work overlaps, but the target is different: instead of ranking a page, the goal is being present and described correctly in the specific sources the models read.
Is this one of those tools that posts from fake accounts? +
No. No fake accounts, no bought personas, no undisclosed posting. Community answers go out under your real name, where they're genuinely useful. Tactics that need hiding get brands banned and look terrible in investor diligence.
How fast do results show? +
Brand-name questions usually move within weeks once your site and profiles are fixed and indexed. Category questions like best X in India take roughly six to twelve weeks, because they depend on third-party sources accepting you. Anyone promising faster is guessing.
Do you guarantee AI will recommend me? +
No, and nobody honestly can, because the models change and the sources decide. What I promise and deliver: the placements, the fixes, and a measured before and after of what each model says. You'll see the movement in screenshots, not in adjectives.
What do you need from me? +
Your website, your three closest rivals, and about fifteen minutes on WhatsApp to understand who your buyer is. That's enough to build the prompt set.
Can I try before paying? +
Yes. Send your website on WhatsApp and a free three-prompt snapshot with screenshots comes back within a day. No payment, no call.
StartupRaise← Every service, one threadValidation to investor outreach, flat fees, over WhatsApp. Next deep pageThe model investors can't break →Four real sheets, live numbers, the 0.00 rule.