KATA AUTOMATION: AI SEARCH VISIBILITY CHECKLIST
Version: 9 September 2026
Guide: https://kataautomation.ai/guides/ai-search-visibility/
Kata's baseline: https://kataautomation.ai/work/seo-ai-search/

YOUR TEST
Business:
Customer type:
Service:
Target market:
Prompt version:
Dates and time zone:
Actual country/location:
Language:
Google device and result types:
Search Console property, reporting dates and filters:
Index status / chosen canonical for each key page:
Web clicks / impressions / available query rows:
Unavailable data and exclusions:

Replace bracketed text BEFORE the first run. Keep the final wording unchanged for comparisons. Your adapted set is a separate test from Kata's.

PROMPT TEMPLATES
B1: Does it know the business?
What does [business name] in [place] do? Cite your sources.

B2: Does it understand the fit?
Is [business name] a suitable [type of provider] for [type of customer]? Explain what you can verify and cite your sources.

D1: Find a provider
Which companies help [type of customer] in [market] with [main problem]? Cite sources for your recommendations.

D2: Solve a specific problem
Who can help [type of customer] set up [needed service] and improve [desired outcome]? Cite sources for your recommendations.

D3: Find the right delivery model
Which [type of provider] build and manage [needed system or service] for [type of customer]? Cite sources for your recommendations.

D4: Ask for advice and a provider
How can [type of customer] solve [specific problem], and which providers can implement the solution? Cite sources for your recommendations.

RUN EACH PROMPT THREE TIMES PER APP, IN NEW CHATS
Do not put B and D prompts in the same conversation. New chats may retain account context. Keep ChatGPT, Gemini, Google AI Mode and any Claude observations separate. Record unavailable tools rather than zero. Select search where available; a selected control alone does not prove execution.

ONE RECORD PER ANSWER (duplicate this block)
Record ID / attempt:
App / displayed model:
Date / time / time zone / actual location / language:
Account context, memory and personalisation observed:
Exact prompt / version:
Search evidence: observed / selected only / uncertain / not selected
Full response saved at:
Correct business named: yes / no / uncertain
Own website linked: yes / no / uncertain
Exact own-site URL(s):
Other source URL(s):
Unverified source chip(s):
Description: accurate / inaccurate / uncertain; reason:
Recommended for buyer task: yes / no / uncertain
Valid, excluded or uncertain run; reason:

COUNT PER APP, KEEPING BRANDED AND UNBRANDED SEPARATE
Set exclusion rules before testing. Wrong prompts or reused chats can be excluded; an answer that leaves your business out cannot. Preserve uncertain and excluded runs with reasons.
B1: attempt 1 _____  attempt 2 _____  attempt 3 _____
B2: attempt 1 _____  attempt 2 _____  attempt 3 _____
D1: attempt 1 _____  attempt 2 _____  attempt 3 _____
D2: attempt 1 _____  attempt 2 _____  attempt 3 _____
D3: attempt 1 _____  attempt 2 _____  attempt 3 _____
D4: attempt 1 _____  attempt 2 _____  attempt 3 _____
Branded valid answers / expected 6:
Unbranded valid answers / expected 12:
Unbranded mentions / valid answers:
Unbranded own-site linked citations / valid answers:
Unbranded recommendations / valid answers:
Uncertain or excluded runs and reasons:

Direct-URL retrieval checks belong in a SEPARATE record; never add them to discovery counts.
A source chip without an inspected URL is not a verified link.
A small prompt sample is not a universal rank or market share.
Google AI-report impressions are a subset of Web totals: do not add them twice.
Missing data is unknown, not zero. Percentage growth from zero is undefined.
Our own searches may affect later search impressions.

CHOOSE AND RECHECK
Finding:
Evidence:
Next useful fix:
What changed and when:
Next review date:
Changed model/account/site/market conditions:
Later comparable observations:
Visits / qualified enquiries (unknown unless actually measured):

KATA'S EXACT ORIGINAL PROMPTS (reference, not your adapted set)
B1: What does Kata Automation, founded by Robert Reid, do? Cite your sources.

B2: Is Kata Automation a suitable AI growth partner for a small service business? Explain what you can verify and cite your sources.

D1: Which agencies help US service businesses use AI to get more leads and win more work? Cite sources for your recommendations.

D2: Who can help a small service business set up automated lead follow-up and improve sales conversion? Cite sources for your recommendations.

D3: Which AI growth agencies build and manage lead generation and sales follow-up systems for service businesses? Cite sources for your recommendations.

D4: How can a service business respond to new enquiries faster, and which providers can implement the system? Cite your sources.

Kata collected its starting observations on 7-8 September 2026. This is a method and worksheet, not evidence of ranking improvement. Keep private account and customer information out of public evidence. Use your own approved tools; this file sends nothing to Kata.