A media report should help a founder decide what to keep funding. Counting placements answers one question. Observing AI citations answers another. Neither, on its own, answers whether the investment produced qualified demand.
I want those layers connected without pretending they are the same metric.
Keep the chain visible
Start with the work: which coverage was earned, what it actually said, and whether it contained evidence relevant to a buyer.
Then observe discovery: which engines retrieved or cited that coverage, for which questions, during which window. Distinguish a link to a source from a brand mention, a recommendation and a correct explanation of the company.
Finally inspect business outcomes: qualified referrals, recorded discovery conversations and the decisions those buyers made. Where attribution is incomplete, keep it incomplete. A convenient multiplier does not turn a citation into revenue.
Make the denominator explicit
A visibility percentage needs a defined question set, included engines, collection cadence and treatment of failed runs. A change in any of those can move the score without a change in the market.
Likewise, a source ranking across a large benchmark is an aggregate. It does not describe the composition of one answer or the media budget needed to influence it. The Machine Relations Index is a source-pattern reference; company-specific decisions still need company-specific observations.
Unclassified sources are another limit. Classify the underlying URLs before deciding that a source type is missing. “Unknown” is a gap in the analysis, not a category to pitch or a percentage to allocate.
Ask what would change the decision
Before buying a tool, name the choice it will inform. Should the company pursue a trade publication? Repair an obsolete source? Clarify its product pages? Keep a campaign running?
Choose an observation that could change that decision, including one that could tell you to stop. If every possible chart leads to buying more of the same service, the measurement is serving the sale.
Use the build-versus-buy boundary to separate useful collection software from judgment the business should own.
The point is not to make PR answer for every sale. It is to understand what earned authority changed, what remains unproven, and which next action deserves the budget.
About the author
Jaxon Parrott founded AuthorityTech in 2018 and coined Machine Relations in 2024. He writes about earned authority, AI discovery and founder judgment.
Continue reading
The Category Ownership Trap
Owning a category definition does not mean owning the buyer journey. How I connect Machine Relations to the questions people ask before they know its name.
When AI Cannot Recognize the Business You Built
A company can have customers and demand while AI search misidentifies it. Fix conflicting identity, missing evidence and weak retrieval before buying more content.
When AI Recommends Your Competitor
When AI recommends a competitor, inspect the question and cited evidence before launching a campaign. Fix the specific identity, proof or retrieval gap you find.