Skip to content

AI visibility audit · free Shortlist Scan

AI visibility audit: is your firm named when a buyer asks AI?

Agency named most often4 of 6
Agency named once1 of 6
What the Scan shows: one buyer question, 6 AI answers, 12 October 2026. See the answers

An AI visibility audit asks ChatGPT, Perplexity, Gemini and Google AI Overviews your buyers’ questions and records whether each answer names your firm, mentions it or leaves it out. Ours, the Shortlist Scan, is free for firms with partners: 20 questions, 3 runs each, against 3 competitors you name.

Get my free Shortlist Scan →Talk about my projectScore yourself in 3 prompts (10 min) ↓

You send
Your domain and your work email. We agree on your market and your 3 competitors by email
You get
The 20 questions, the answers we captured and your place next to 3 competitors
Reply
A person, by email. Nothing is published, nothing is signed.
To: the Jungle Roots teamJR/07
https://

By submitting, you agree that Jungle Roots can write to you about this request. Your email is used for this Scan and the follow-up about it, and we never sell lead lists.

Definitions real answers, dated

Named, mentioned or absent.

Named
The answer recommends your firm by name. The only state that puts you on a buyer’s shortlist.
Mentioned
Your firm appears, but not as a pick: in passing, as a caveat, or only as a linked source under the answer.
Absent
Neither your name nor your pages appear. The buyer never meets you through that answer.

In the second Perplexity answer of 12 October 2026, Agency D, Agency F, Agency E were named; Agency G, Agency H, Agency I were mentioned as sources only.

Named, mentioned or absent: one question, six AI answers, 12 October 2026Grid of 6 agencies by 6 answers. Named counts: Agency D 4, Agency E 4, Agency F 3, Agency G 2, Agency H 2, Agency I 0.named: the answer recommends your firm by namementioned: your firm appears, but not as a pick: in passing, as a caveat, or only as a linked source under theanswerabsent: neither your name nor your pages appearCHATGPT, 3 ANSWERSPERPLEXITY, 3 ANSWERS#1#2#3#1#2#3NAMEDAgency D4 of 6Agency E4 of 6Agency F3 of 6Agency G2 of 6Agency H2 of 6Agency I0 of 6Also named once each, all by ChatGPT: Agency A, Agency B, Agency C, Agency J, Agency K, Agency L, Agency M, Agency N,Agency O.Named, mentioned or absent: one question, six AI answers, 12 October 2026Grid of 6 agencies by 6 answers. Named counts: Agency D 4, Agency E 4, Agency F 3, Agency G 2, Agency H 2, Agency I 0.named: the answer recommends your firmby namementioned: your firm appears, but notas a pick: in passing, as a caveat, oronly as a linked source under theanswerabsent: neither your name nor yourpages appearCHATGPTPERPLEXITYAgency D4/6Agency E4/6Agency F3/6Agency G2/6Agency H2/6Agency I0/6
What the figure shows. One buyer question, “Which agency can help an executive search firm get recommended by ChatGPT and Perplexity?”, asked 3 times to ChatGPT and 3 times to Perplexity on 12 October 2026. Agencies D and E were named in 4 of 6 answers, Agency F in 3. Agencies G and H were named twice and, in one Perplexity answer, appeared only as a linked source: mentioned, not recommended. Agency I was only ever a linked source. Nine other agencies were named once each, all in ChatGPT, and ChatGPT put a different agency first in each of its 3 answers. Names are hidden here; the capture file keeps them. The conclusion: one screenshot proves nothing. Count how often a firm is named across repeated answers, and separate “named” from “linked”. Raw answers: capture file.
Do it yourself 10 minutes

Run a 10-minute self-check first.

Open ChatGPT in a private window and ask prompts A, B and C three times each, with your own practice, firm and city in the brackets. Keep the best of the three runs: named (the answer recommends your firm by name), mentioned (your firm appears, but not as a pick: in passing, as a caveat, or only as a linked source under the answer) or absent.

Self-check: score three answers

1. What kind of business are you?
2. What does your website publish today?
Prompt A: “Which [your practice] firm should I call for [your service] in [your city]?”
Prompt B: “What does [your firm] do, and who are its partners?”
Prompt C: “Compare [your firm] with [your closest competitor].”

Example: how a 2/6 score reads

A worked example, built to show how the score reads. Your own answers start blank on the left.

2/6 Named on Prompt B; absent on Prompt A and Prompt C.

Fix Prompt A first: no page or independent source answers the category question. Publish one page per service that answers it, and get listed where your profession ranks firms.

Scoring: named 2 points, mentioned 1, absent 0, out of 2 per prompt tested.

  • Prompt A, “Which executive search firm should we call to hire a fintech CFO in Paris?”: if the firm is left out of all 3 answers, the prompt scores 0.
  • Prompt B, “What does [the firm] do, and who are its partners?”: if the firm is named and described correctly, the prompt scores 2.
  • Prompt C, “Compare [the firm] with [a competitor].”: if the answer leaves the firm out, or mixes it up with a namesake, the prompt scores 0.
Method

How the Scan is run.

  1. 1

    The questions

    We draft 20 questions from your service pages, your partners’ sectors and the 3 competitors you name: 8 that ask which firm to call, 6 about a problem you solve, 4 comparisons and 2 about your firm by name. You approve the list before we run it.

  2. 2

    The runs

    Each question goes to ChatGPT, Perplexity, Gemini and Google AI Overviews, 3 times, in fresh sessions. Answers vary between runs, so a firm counts as named on a question when at least 2 of the 3 runs name it, and the report shows the count.

  3. 3

    The reading

    An agent captures the answers and their sources. A person on the Jungle Roots team, led by Loïc Guyon, reads every answer and classifies it before anything reaches you. Raw answers come with the report.

  4. 4

    What you get

    The 20 questions, the answers, your state next to the 3 competitors on each one, and the missing sources that explain the absences, in the order worth fixing. We reply by email; nothing is published and nothing is signed.

If the gap is confirmed

Three paid ways to go further. Or stop at the Scan: it is yours.

Shortlist Report

$499 once, excl. tax

For a partner who wants the gaps in writing before deciding anything.

  • A written report on where your firm stands on the shortlist
  • The missing sources behind each absence, in the order to fix them
  • Prepared by our agent, reviewed by a person
  • No meeting
Request the Report →

Shortlist Session

$999 once, excl. tax

For partners who want to question the findings with a person before choosing.

  • The same written report
  • Worked through live by a human consultant with the agent
  • Your partners’ questions answered in the session
  • What changes from the Report: a person, live
Request the Session →

Shortlist Diagnostic

$2,999 once, excl. tax

For a firm ready to compare packs on evidence.

  • 60 questions across 4 engines, 5 named competitors
  • Entity and proof audit
  • 90-day and 12-month plans
  • Readout to the partners, then 30 days of Q&A; about three weeks
Request the Diagnostic →

A pack is our monthly retainer ($1,300 to $7,900 a month, initial term 3 to 6 months) that works your site, your partners, the press and the AI answers; the Report and the Session are deducted if you sign, the Diagnostic from your first monthly fee if you sign.

See every price →Each paid step is deducted if you sign a pack.

Questions about AI visibility audits

What is an AI visibility audit?

A structured check of how AI engines answer your buyers’ questions: which firms they name, which sources they cite, and whether your firm is named, mentioned or absent. It is run on a fixed list of questions so it can be repeated and compared.

How is it different from an SEO audit?

An SEO audit checks how your pages rank on a results page. An AI visibility audit checks what the generated answer says, which firms it names and which pages it relies on. A firm can rank well and still be absent from the answer.

Why ask each question three times?

Because answers change between runs. On 12 October 2026, ChatGPT put a different agency first in each of three answers to the same question. Counting how often a firm is named across runs separates a steady position from a lucky one.

Who runs the Shortlist Scan?

We write the 20 questions from your pages, your partners’ sectors and the 3 competitors you name, and you approve the list. Each question goes 3 times to each engine, in fresh sessions. An agent captures the answers and their sources, then a person on the Jungle Roots team reads and classifies every answer before the report goes out. The raw answers come attached to the report.

What does an AI visibility audit cost?

The Shortlist Scan is free. The Report costs $499 and the Session $999, each deducted if you sign a pack. The Diagnostic costs $2,999, takes about three weeks and is deducted from your first monthly fee if you sign.

One audit, one action plan, and you choose who applies it.

Every page on this site leads to the same starting point: the free Shortlist Scan, then the Diagnostic. The action plan that comes out of it is built for your firm, channel by channel, and applied with you or for you.

Run the Shortlist Scan