How should you measure and compare AI visibility?
Broadcastwell measures and fixes whether AI search engines name B2B software companies.
Run the free checkCompare AI visibility reports using counts, citations, dates and repeat runs. Each outcome needs a defined event and base before the rates can be compared.
Observed answers on stated dates, not a ranking.
These 47 answers use dated Broadcastwell findings. Each page identifies its question, figure and source; a count in an answer does not establish product quality or a completed sale.
Questions
What is a good AI visibility score?
No universal good AI visibility score follows from our data. Compare the question set, category and valid base before interpreting a company's naming rate.
How do we compare citation metrics when changing AI tools?
Match definitions before rates: Housecall Pro was named in 108 of 150 answers and its own domain cited in 43 of 150.
Reports and evidence
What should a client AI visibility report contain?
A client report needs event, base, date and receipt: Housecall Pro was named in 108 of 150 answers; 43 cited its domain.
What evidence can connect AI visibility with pipeline?
Keep dated answer, visit, buyer-account and opportunity records, with a rule for joining them. Missing connections remain unknown rather than attribution.
Historical comparisons
How can I check a tool's historical AI visibility data?
Check the records behind a history chart: HubSpot 77/117 in August and 113/150 in September came from different collections.
Counting answers
What counts as a brand mention in an AI answer?
A brand mention is a matched name in an answer. Housecall Pro was named in 108 of 150 answers on 22 September 2026; that is not a recommendation score.
What counts as an own-domain citation in an AI answer?
In the September Index, a citation requires a matched vendor name plus a supported vendor domain. Housecall Pro met both in 43 of 150 answers.
Search Console evidence
What can Search Console show about AI visibility queries?
Search Console reports query-table activity: a dated length filter found 134 distinct queries, 969 impressions and zero query-table clicks.
Source observations
Does Medium get cited in AI answers?
Yes. Medium was cited in six of 150 project-management answers in September 2026, with other counts in CRM and hotel panels. These are exact-host observations.
Baseline design
How should I check an AI baseline before changing content?
Fix the question set, retain answers and check repeat variation first: Claude’s mean shortlist agreement was 0.442 over 74 pairs on 4 August 2026.
Repeat counts
How many repeated runs make an AI visibility rate useful?
There is no universal sufficient run count: Procore’s 3/3 AI Mode result on 22 September 2026 has a descriptive 95% interval of 43.9% to 100.0%.
Roadmap decisions
Which AI visibility signals should shape an SEO roadmap?
Investigate missing names, missing citations, unsupported facts and unstable results separately. A combined visibility score can hide the task that needs work.
Claim support
How can I check whether an AI citation supports its claim?
Read the cited passage against the answer's subject, date, number and conditions. A citation label alone cannot establish that a source supports the claim.
Visits and answers
What does a recorded AI referral establish?
A recorded AI referral is a visit classification, not an answer count or sale. Housecall Pro’s 108/150 names and 43/150 citations were answer observations.
Checking tools
Which AI visibility tools can a buyer verify?
Check Profound and seoClarity against answer receipts: Jobber was named in 50 of 150 answers and its domain cited in 46 on 22 September 2026.
How can I evaluate Hotwire GAIO.tech's AI visibility work?
Hotwire called Spark the successor name for GAIO.tech on 27 March 2025. Start there, then check the product's scope and the answer events its reports count.
Tool accuracy evidence
What does Hotwire GAIO.tech publish about data accuracy?
Hotwire's GAIO.tech and Spark announcements describe visibility and sources. The two pages read on 29 Oct 2026 supplied no claim-by-claim accuracy result.
Separate client reporting
Which AI visibility platforms support agency clients?
OtterlyAI documents separate client Workspaces. Jobber’s 50 naming and 46 own-domain-cited answers illustrate why each client report needs exact event labels.
Comparing measured events
Which measures belong beside one another?
Procore was named in 109 of 150 answers on 22 September 2026; 30 also cited its domain. Keep both event counts distinct; neither is an overall success score.
Athena accuracy definitions
What does Athena publish about data accuracy?
AthenaHQ's FAQ describes answer collection and visibility tracking. Customer outcomes do not replace the claim and reference needed for a truth check.
Dated Hotwire engine labels
What does Hotwire GAIO.tech publish about model coverage?
Hotwire named four chatbots in its GAIO.tech launch and a different list in the Spark announcement. Read the dated labels separately from our engine counts.
Provider sampling definitions
How do AI trackers sample answers and count share of voice?
Compare sampling rules and denominators across Suede, Profound and AthenaHQ. Brand mentions can occur in one answer, so mention totals and answer totals differ.
Named platform history features
Which AI visibility platforms offer historical data and time-series reporting?
Otterly documents daily prompt history; Profound documents daily answer datasets and charts. Read the run identity behind each trend.
Hotwire’s product and service names
What AI visibility products does Hotwire GAIO.tech offer?
Hotwire's 27 March 2025 announcement calls Spark the former GAIO.tech visibility tool. It describes Ignite separately as an account-intelligence product.
Documented audit scope
Does Profound provide a full SEO audit or a partial audit?
Profound documents answer presence, citations and brand-claim checks. Those features do not establish the scope or completion of a technical SEO audit.
Checking a saved answer sample
How can I check AI search visibility?
Check the exact question, engine and saved answer. Count names and citations separately from rankings, visits and sales.
The Index’s checked panel and units
What is an AI visibility index?
An AI visibility index records what a set of answers names or cites. September 2026's Absence Index covers 286 vendor/category records across 14 categories.
Checking collection freshness
How can I check whether an AI tracker uses live answers?
Inspect receipts, saved responses and cache settings to check a tracker's live-answer claim. A dashboard refresh does not prove a new answer was saved.
Threads access and recorded citations
Can Google, Bing and LLMs find content on Threads?
A public Threads URL does not prove search or AI visibility. No Threads host appeared in 2,093 included Answer RFP responses from 22 and 23 September 2026.
Claude’s answer and citation records
How can I track brand visibility in Claude?
Keep each Claude question and answer, then count the checked brand under a stated alias rule. Counts across brands cannot serve as one brand's answer share.
AIO website tracking
Which tools track website presence in Google AI Overviews?
seoClarity describes AI Overview query filters, linked-domain details and SERP Previews. Inspect the answer and destination URL behind a site appearance count.
Lighthouse file retrieval and citation records
What does Lighthouse’s llms.txt audit check?
Lighthouse describes an optional llms.txt retrieval check under Agentic browsing. Passing that check does not prove indexing, citations or a visibility effect.
Company-fact errors
Can AI answers get company facts wrong?
Our dated Open Dental record labels 12/77 statements wrong or stale. Check the claim class, capture date and retained vendor reference.
Documented ChatGPT GEO cases
Who has documented results from GEO work in ChatGPT?
Suede publishes a dated Koomba ChatGPT case using a named-or-cited rate. Our Close 3/30 example distinguishes a naming-only observation.
Answer-audit scope
How does an AEO or GEO audit differ from an SEO audit?
An answer audit inspects retained claims as well as site eligibility. Our construction study found a pricing criterion in 5/149 included answers.
GEO effectiveness
Is GEO more effective than SEO?
Our evidence does not establish that GEO is more effective than SEO. The September 2026 Index counts answers, without comparing traffic, sales or returns.
Reddit citations
Do AI product answers cite Reddit?
Yes. Google AI Overviews cited Reddit in 23 of 30 radiology answers on 22 Sep 2026; ChatGPT cited it in none of 30. Name the engine and category for each count.
AI Mode and site clicks
How does Google AI Mode affect site clicks?
Our Mode CRM panel has three HubSpot-domain citing answers. Their retained URLs and Google’s Web report distinguish source selection from site clicks.
Google rank and answer-source selection
Why can a page rank in Google without being cited by ChatGPT or Perplexity?
Salesforce-domain URLs appeared in 19 ChatGPT and three Perplexity CRM answers. Compare those source selections with the services’ eligibility rules.
Observed crawler requests to llms.txt
How can logs show which AI crawlers requested llms.txt?
A server request needs a path, time, status and a defensible crawler identity. Our zero llms.txt citations in 2,093 answers measure a different event.
YouTube citations and transcript evidence
Do YouTube transcripts affect LLM citations?
PHP2Shell's 23 Oct 2026 collection found YouTube URLs in 21 of 148 eligible answers. Those citations do not establish whether transcripts affected the answers.
Verifying a reported conclusion
How can a buyer verify the conclusion?
The 12/77 Open Dental finding has dated claims and vendor references. Use the statement, base, status and source pair to check the historical conclusion.
Presenting separate outcome records
How should a report present these outcomes?
Report answer observations, visits, buyer accounts and commercial events separately. Retain each measure's date, event definition, evidence and applicable base.
Visit definitions and report windows
How should the report show visits?
A visit report needs its period, visit definition and source evidence. An AI answer naming a company cannot supply the website's visit count or traffic source.
Source and business evidence
A buyer report is a buyer's account of finding or assessing you. An assistant selecting a vendor does not establish what a person said, clicked or bought.
What does the business record establish?
Use a dated business record to establish an enquiry or sale and its status. Words in an AI software answer cannot supply evidence that either event occurred.
What information belongs in a public report?
Publish reviewed aggregates and authorised evidence with dates, bases and sources. A public study count does not permit publishing private client or deal notes.
Browse all answer clusters. For the full argument and measurement rules, read the Absence Manual.
Prepared by the Broadcastwell team.
Page owner: Vikram Patel, AI Architect. Corrections: hello@broadcastwell.com
Related research
- Why does a search result need a date and configuration?
- What do Rovira's search results say about citations?
- Do university website scores explain academic standing?
- What does Share of Model measure?
- What is citation decay?
Published . Last updated .