MMarketing Against The Grain
← All frameworks
Marketing

AI Visibility Value Model

Estimate AI-search impact through impressions, conversion quality, and speed

Difficulty
Moderate
Time to result
~weeks to results
Steps
6
Confidence
97%

The AI Visibility Value Model evaluates answer-engine exposure as a mixture of brand marketing and high-intent acquisition rather than a conventional click channel. Begin with direct referrals from ChatGPT, Gemini, and similar systems, but treat them as a lower bound. Estimate a range of unseen impressions because many answers mention a brand without a citation click, and users may later visit directly, search the brand, or respond to another touchpoint. Then compare the conversion rate and sales-cycle speed of identified AI-assisted visitors against conventional search traffic. Supplement analytics by searching call recordings for customers who mention using an answer engine during research. The model deliberately reports ranges and correlations rather than exact attribution. Its purpose is to estimate strategic value without pretending that incomplete click data captures every exposure or assisted purchase.

Origin

Extracted from Marketing Against The Grain as the hosts explained why Hampton's recorded ChatGPT referrals understated its AI-search impact.

Core principles

  • 01Referral clicks reveal only a fraction of AI influence
  • 02Most value may come from unclicked mentions and later visits
  • 03AI-assisted visitors can convert at a higher rate
  • 04AI conversations compress research and consideration time

How to run it

  1. 1

    Establish direct referrals

    Measure identifiable visits from relevant answer engines over a consistent period.

    Pro tip Separate engines where analytics permits.

    Watch out Referral traffic is a floor, not the complete exposure count.

  2. 2

    Estimate unseen impressions

    Apply a defensible range derived from available platform or advertising data to approximate mentions that did not produce clicks.

    Pro tip Use scenario ranges and update them when better impression data becomes available.

    Watch out Do not present a rough multiplier as an audited organic impression count.

  3. 3

    Capture indirect discovery

    Monitor direct visits, branded search, surveys, and call transcripts for evidence that an AI answer influenced later behavior.

    Pro tip Ask customers how they first encountered and researched the brand.

    Watch out Avoid assigning every increase in direct traffic to AI search.

  4. 4

    Compare conversion quality

    Calculate conversion or customer rates for identifiable AI referrals and compare them with conventional search sources.

    Pro tip Segment by similar intent and time period where possible.

    Watch out Small samples and selection effects can exaggerate apparent lift.

  5. 5

    Measure purchase speed

    Compare time from first known interaction to purchase for AI-assisted and other customers.

    Pro tip Use CRM and call data to understand whether research occurred before the tracked visit.

    Watch out A short recorded sales cycle may omit earlier anonymous research.

  6. 6

    Make a range-based decision

    Combine exposure, conversion quality, sales-cycle speed, and customer testimony into a bounded investment case.

    Pro tip State assumptions explicitly and revisit them after a test period.

    Watch out Do not wait for perfect attribution if customer behavior is visibly shifting, but do not overclaim causality.

In the wild

Hampton's 1,200 ChatGPT visitors

Hampton identified roughly 1,200 ChatGPT referral visitors over three months. The discussion treated these as people who clicked citation links, while recognizing additional unlinked mentions, direct visits, later branded searches, and remembered recommendations.

The measured referrals became a lower-bound indicator rather than the complete estimate of AI influence.

HubSpot's conversion comparison

HubSpot compared identifiable AI-search visitors with traditional blue-link search visitors and also searched sales-call transcripts for customers who described AI-assisted research.

The company assessed traffic replacement through conversion quality and sales velocity, not raw visit volume alone.

Common mistakes

Counting only referral clicks

Users may see an unlinked mention, visit directly, search later, or remember the recommendation, none of which appears as an answer-engine referral.

Claiming exact impressions

Organic impression data is incomplete, so multipliers should be presented as estimates or ranges with their source and assumptions.

Ignoring customer quality

Raw traffic comparisons miss whether AI-assisted visitors convert more often or complete the buying process faster.

Is it for you?

Best for

Businesses deciding whether AI-search visibility deserves budget despite incomplete organic impression data.

Not ideal for

Organizations requiring exact user-level attribution before recognizing brand or assisted-conversion effects.

From the transcript

AI search is much more akin to like brand marketing, where you're just trying to get massive amounts of impressions versus performance marketing where you…

Kieran · 19:00

for however many clicks you've already gotten, assume that there's a hundred times more pressure.

Sam Parr · 33:30

the people who come to us from AI search directly become customers at like three to five X rate, but they also become customers like…

Kipp · 37:00

From the episode

The New Rules to Beat the SEO Algorithm in 2026