Adobe's 261 million prompt claim needs a query-population and attribution ledger

Define the population behind every visibility rate

A visibility percentage has no stable meaning until marketing records what could have been observed. Define the audience and decision, seed topics, prompt-generation method, real versus synthetic prompts, languages, geographies, devices, model products, account states, dates, repetition schedule, and exclusions. Preserve the exact prompt and returned answer rather than only a normalized topic. Record whether the brand appeared in answer text, a cited source, a link, a shopping element, or an inferred entity, and distinguish owned, earned, partner, retailer, and unauthorized sources. If provider coverage changes, maintain a bridge so an apparent gain is not merely a larger or different sample. The CMO should reject a single blended score when its denominator and observation windows cannot be reproduced.

Separate answer presence from attributable demand

Presence, citation, referral, and conversion are different events. Create a ledger that connects an observed prompt and answer to the cited or linked destination, referral session where observable, consent state, campaign and content identifiers, downstream action, and attribution rule. Document where the chain breaks because an assistant does not expose a referral, a user later navigates directly, a platform strips detail, or cross-device identity is unavailable. Compare against an appropriate baseline and preserve unattributed outcomes. A correlation between higher AI visibility and stronger traffic or conversion does not prove the visibility caused the behavior; brand demand, media, seasonality, distribution, and content changes may affect both. Use the metric for diagnosis before using it for budget claims.

Audit the measurement system for drift and gaming

Run repeated tests on a frozen query set and a time-varying discovery set. Measure answer variance, missing observations, model and region coverage, citation matching, entity ambiguity, duplicate prompts, referral classification, and reconciliation with first-party analytics. Keep screenshots or response records where terms permit, method versions, source timestamps, and correction history. Test misleading edge cases: a citation that contradicts the brand, a reseller page mistaken for an owned source, an answer that mentions a namesake, a prompt repeated across minor variants, and traffic generated by automated agents. Require change notices for the provider's prompt corpus, collection methods, scoring, integrations, or models. A metric that can move without an explainable market or content event needs investigation, not a celebratory narrative.

Tie the dashboard to reversible marketing decisions

Approve limited uses first: identifying recurring factual gaps, prioritizing source corrections, comparing citation eligibility for a defined topic, or testing whether an authoritative page becomes more consistently represented. Name the content, analytics, legal, privacy, brand, and channel owners. Establish thresholds for action, prohibited manipulation, review frequency, and stop conditions if coverage or attribution deteriorates. Do not infer search demand, market share, brand preference, incremental revenue, or universal assistant performance from the provider's aggregate prompt count. A budget change requires reconciled first-party outcomes and a counterfactual appropriate to the decision. Human owners remain accountable for claims, source quality, audience impact, and brand actions even when the monitoring platform automates collection and scoring.

Turn this source into a reviewable decision

For AI for CMOs, use this briefing as a dated decision record rather than a substitute for the source. Preserve Adobe, the exact URL, the August 28, 2026 review date, the supported facts above, the editorial interpretation, the limitations, and any buyer-specific evidence. Link that record to the decisions most directly affected: Audience and market insight; Content supply-chain operations; Measurement and performance explanation; Brand, disclosure, and synthetic-media risk. State whether the source changes the scope, evidence requirement, control, sequence, or only the language used to describe the decision.

Before action, name the accountable owner, affected population and workflow, exact offering or configuration, source data and rights, human decision point, exception and appeal path, complete cost, expected benefit, failure and stop conditions, retained evidence, and next review date. Keep official facts, provider statements, buyer observations, representative tests, measured outcomes, editorial inferences, and unknowns visibly separate. Reopen the record when the source, offer, model, integration, data, policy, population, responsible person, or measured result changes.

Limitations and unknowns

Adobe is the provider source. Its June 17, 2026 announcement describes Adobe Brand Visibility and claims a data foundation that includes 261 million real AI-search prompts, CDN-observed agentic traffic, and LLM referral traffic connected through Adobe Analytics. It does not independently establish corpus representativeness, model and geography coverage, collection consent, observation failures, scoring and deduplication rules, entity resolution, referral completeness, attribution method, causal incrementality, buyer configuration, total cost, or business outcome. Current product and contract documentation, buyer-specific measurement design, method-change records, representative tests, reconciled first-party analytics, and qualified analytics, marketing, brand, legal, privacy, security, procurement, accessibility, and finance review control.