Why it matters.
Forrester reported on 21 January 2026 that 94% of business buyers use AI during their buying process, and that 13 internal stakeholders and nine external participants influence a typical decision. Many of the people who see your name in an AI answer will never visit your site.
So no single metric covers the effect. Blend the layers and you hide problems: more AI mentions with flat pipeline, or a jump in enquiries that turns out to be spam. Kept apart, each layer explains the next.
The 10-minute self-check.
Open your last marketing report. Can a reader tell, separately and with a date range for each:
- How often the AI engines name you for your buyers’ questions?
- How many visits and key events came from AI assistants and from brand searches?
- How much pipeline and closed revenue those sources produced?
Steps.
- Visibility layer: share of runs that name you, per engine, from a fixed question panel (plays 02 to 04).
- Behaviour layer: AI Assistant sessions, branded search, direct visits and key events (plays 36 and 38).
- Pipeline layer: enquiries by self-reported and tracked source, opportunities and closed-won value (plays 37 and 40).
- Put a source and a date range on every number. Missing data is written “not captured”, never estimated.
- Keep the layers side by side; never add them into one score.
- Close with three lines: what changed, why we think so, what we do next.
This is the reporting model we use in our conversion tracking and analytics work.
What we saw on prolinkage.com.
How to measure it.
- Visibility: named rate per engine, monthly.
- Behaviour: AI and branded sessions and their key events, monthly.
- Pipeline: opportunities and closed-won value by source, quarterly, because deals take longer than a month.
Sources.
- Forrester, The State of Business Buying, 2026, 21 January 2026.
- Prolinkage learnings log, lead replay, 6 November 2026.
- Prolinkage change ledger, GA4 fixes (property 399997904), 6 November 2026.