Question source and wording
Save why the question matters and keep the exact words.
AI Visibility · Measurement
An AI mention is useful evidence. It is not revenue.
Save the exact question, tool, date and answer. Keep visits, enquiries and completed business in separate records.
Six separate records
Start with the source of the number.
First ask where the number came from. Was it a search platform, an analytics tool, a saved set of AI tests or a business system? Each source answers a different question.
Mindflow keeps these records apart. A saved AI answer should never turn into an unsupported claim about reach, leads or revenue.
| Layer | Record | What it establishes | What it does not establish |
|---|---|---|---|
| Answer observation | Exact retained output | What one logged run showed | Audience reach, permanence or buyer influence |
| Official platform record | Impression, citation, page or query field | What the platform reported | The full buyer journey or choice |
| Visit | Referral session and landing page | A trackable click path reached the site | Qualification or a sale |
| Onsite action | Defined analytics event | A tracked action took place | Lead quality or business value |
| Qualified action | Agreed business rules | The action met the rule | A completed sale or result |
| Commercial outcome | Scheduling, CRM or store record | The business recorded the result | Proof that one AI answer caused it |
Save the exact check
Save the test details before you score the answer.
AI answers can change with the question, tool, mode, account, location, language and time. A screenshot without those details is hard to compare.
Give each test its own ID. Save the full answer, not just the line that names the business. Also save the test method and the date for the next check.
Tool and mode
Name the product and mode. Do not mix results from different tools.
Account and location
Save the account state, place, language and device when known.
Timestamp and version
Capture the date and any exposed model or interface version.
Full answer
Keep the full reply, visible sources, links and answer position.
Method and next check
Save the math, test-set version, notes and rules for the next run.
Different answer signals
Keep mentions, recommendations and citations separate.
A mention names the business. A recommendation offers it as a choice. A citation shows a source. One answer may contain one, two or all three.
Position, accuracy and tone are separate. A large but wrong mention is not good evidence. A citation may support a fact without naming the business.
The business was named
Useful for presence in the sampled output. It does not establish endorsement, source use or a visit.
The business was framed as an option
Keep the exact wording and qualification. It is not a stable rank or a future promise.
A source was exposed
Record the cited page and relationship. The citation does not prove a click or buyer trust.
The facts were checked
Compare material names, services, products, people, locations and limits with current sources.
Show the denominator
Put the calculation beside every score.
A count or percentage can be useful. The problem starts when the report hides the total number of tests, the source questions, the companies compared or the scoring rule.
“7 of 12 saved tests” describes that sample. It is not audience share, market share or the chance that a buyer will see the business. More runs can show change, but no fixed number makes the answer certain.
Before Mindflow publishes one visibility score, the record must name the saved questions, platforms, number of runs, dates and formula. It must also explain how mentions, recommendations and citations count, how missing answers are handled, what uncertainty is shown and when the score will be updated. Until that operating record exists, use disclosed sample counts instead of a headline score.
Mindflow shows the total beside every sample score. Sales and revenue stay in a separate record.
| Metric | Required denominator or method | Safe interpretation | Do not call it |
|---|---|---|---|
| Mention count | Logged runs in the disclosed panel | Presence within that sample | Total audience reach |
| Recommendation rate | Recommendation runs ÷ eligible logged runs | Sample frequency under those conditions | Conversion probability |
| Citation share | Citations, source set and comparison scope | Relative source appearance in the sample | Universal authority |
| Proxy impressions | Vendor model and external search-volume inputs | A constructed estimate | Actual platform impressions |
| Competitor share | Named competitor set and weighting | Comparison inside the selected panel | Whole-market share |
Keep X/12 as a disclosed sample—not the headline business result.
First-party platform records
Use the platform’s own data when it is available.
Official reports reduce guesswork because the platform defines the number. The data can still be limited or grouped.
Outside tools can help run saved AI tests. They do not have secret access to a platform’s ranking or AI systems.
Mindflow uses the official number first. Saved test sets fill only the gaps the platform does not report.
AI impressions
Google’s report can show impressions by page, country, device and date for some sites. The data is limited. It does not show every question or lead.
Citations and cited pages
Bing reports citations and sample questions. A citation count does not show where the link appeared or how much it mattered.
Citations and AI visits
Clarity shows citations, cited pages and AI-referred visits as separate numbers. It does not prove that one caused the other.
Trackable clicks
OpenAI adds a source tag to ChatGPT search links. Analytics may record that click. It cannot show influence when the buyer never clicked.
Actions and assigned credit
The business chooses which site actions to track. An attribution model assigns credit. It does not prove cause.
Platform sources checked 16 August 2026. AI reporting fields are changing quickly and require a current pre-publication review.
After the answer
A visit is not yet a good lead or sale.
A referral session says a tracked visit reached the site. A site event says a tracked action took place. Neither tells you if the request fit, the appointment was right or the order stayed profitable.
Match the visit to the business record. Keep analytics credit, human review and the final result separate.
Do not borrow another company’s AI referral conversion rate. Before comparing channels, define one comparison group by referral source, landing page, buyer intent, tracked action, qualification rule, final business result, comparison channel, time period and sample size. The result describes that group, not all AI visitors.
Referral session
Record source, landing page, date and available campaign fields.
Onsite action
Record the call click, form, booking request, demo, trial, cart or checkout event.
Fit check
Apply the business’s service, patient, case, buyer or order rules.
Business record
Confirm the result in the job, booking, intake, CRM or store system.
Five outcome records
Use the right business result for each industry.
The AI test can stay the same. The good action, privacy rule and final business record change.
| Industry | Qualified action | Commercial outcome | Required boundary |
|---|---|---|---|
| Home Services | Valid service request | Booked or completed job | Match calls and forms to the job record |
| Healthcare | Right appointment request | Scheduled appointment | Protect privacy and use the booking record |
| Legal | Possible-client consultation request | Accepted or signed matter where allowed | Use reviewed intake, local rules and privacy controls |
| B2B SaaS | Demo or trial that meets the rule | Good sales opportunity or pipeline | Make product and CRM terms match |
| eCommerce | Product, cart or checkout action | Order, return and repeat purchase | Use the store system as the final record |
Claim-confidence key
Say how strong each claim is.
Mark each finding as observed, reported, inferred, attributed or unknown. The label shows what the evidence can support.
| Label | Use it when | Example | Boundary |
|---|---|---|---|
| Observed | Mindflow saved the dated answer | The business appeared in 7 of 12 saved tests | The finding belongs to that sample |
| Reported | An official platform showed the number | Search Console reported a set number of AI impressions | The platform’s definition controls |
| Inferred | Records moved together but cause is unresolved | Branded visits rose during the same period | Timing alone does not prove cause |
| Attributed | A stated model assigned credit | Analytics gave last-click credit to a ChatGPT visit | Name the credit model |
| Unknown | Evidence is missing or incompatible | No-click influence cannot be reconstructed | Do not fill the gap with a score |
Decision-ready reporting
End every report with one clear decision.
A useful report says what changed, how sure the team is, who owns the next step and when to check again. It also says what new fact would change the decision.
Choose one next step. Correct a wrong source, improve a page, repair access, change the test, fix the handoff or wait for more evidence.
Keep the raw evidence beside the summary. A reader should be able to trace a headline number to the saved test, platform field, visit or business rule behind it. Show it when two systems use different dates or terms.
“No action” can be the right answer when the change is small or unstable. Waiting for a better sample may be safer than changing a useful page after one noisy test.
Name the movement
State the source, period, baseline and exact field that changed.
Name the evidence label
Use observed, reported, inferred, attributed or unknown.
Name the owner and due date
Assign the smallest useful next step.
Freeze the next conditions
State when, where and how the comparison will run.
Say what could change the decision
Name the missing fact or result that would point elsewhere.
See how evidence becomes a Mindflow priority → Browse the research record for the standards used to separate a source, an observation and a Mindflow conclusion.
Review what each number means
Know which numbers are official, sampled or tied to sales.
Bring one AI visibility report or starting point. Mindflow can show which numbers come from a platform, saved answer, site action, fit check or business result. It can also show which claims go too far.
The review does not promise citations, higher scores, leads or revenue credit. Healthcare and Legal work still needs the right privacy, clinical, local-law and intake review.
Primary platform and measurement sources checked 16 August 2026. Re-check within 30 days of publication.