Why AI attribution still has no shared measurement standard

In the space of a week, three developments showed how differently companies are approaching the commercial impact of AI search.

Peec launched AI Referrals on 11 September, connecting AI-assistant traffic with GA4 sessions, conversions and revenue. Semrush published new manufacturing-sector data on 8 September. Then, on 14 September, Estée Lauder announced a global partnership with Profound to manage how its brands appear across AI platforms.

Add the attribution model AirOps has been developing through Page360, and the lack of a shared standard becomes obvious. All four companies are trying to connect AI visibility with business value. They are measuring different parts of that relationship.

If you haven't already covered the basics of getting cited by AI search, that comes before this stage.

What each company is actually measuring

Peec focuses on the traffic that can be seen. Its GA4 integration connects AI referrals with sessions, engagement, conversions and revenue across ChatGPT, Claude, Gemini, Perplexity and Copilot.

Peec also acknowledges the limit of that data. Google classifies clicks from AI Overviews as ordinary Organic Search, while visits from AI apps without a referrer header can appear as Direct. The integration can improve the reporting of identifiable AI traffic, but it cannot reveal every journey influenced by AI.

Semrush's research shows why referral traffic only tells part of the story. Across manufacturing and industrial search, AI Overviews grew from 38% to 57% of tracked search volume between January and July 2026. Yet AI assistants and AI Mode generated just 0.48% of website sessions in the sector over the same period.

The gap between visibility and traffic was just as striking. Of the top 15 brands for AI mentions and citations, only two also appeared in the top 15 for referral traffic. A brand can be highly visible in AI answers without seeing a corresponding rise in identifiable visits.

AirOps takes a broader attribution view. Page360 brings AI citations and mentions together with GA4 and Google Search Console data, while its recommended attribution model also uses CRM and self-reported sources to capture influence that last-click reporting misses.

AirOps cites SegmentStream analysis across its customer base in which AI-attributed revenue rose from 2% under last-click attribution to 16% when identity and self-reported data were included. This is an aggregate vendor result, not a universal benchmark. It does, however, show how much the answer can change when the measurement method changes.

Profound sits further upstream. Its partnership with Estée Lauder will examine how brands including Clinique, M·A·C and La Mer are represented and recommended across AI platforms. Those insights will then inform content across product pages, blogs and YouTube.

The immediate priority is improving the information AI systems use before making a recommendation, not proving that an AI interaction caused a sale.

Why the numbers tell different stories

AI visibility has become an umbrella term for several distinct outcomes.

A mention shows that the brand appears in an answer. A recommendation shows that the AI has selected it as an appropriate option. A citation shows that a particular source was used as evidence. A referral shows that somebody clicked through. Conversion shows whether the visit produced a commercial result.

These signals are related, but they are not interchangeable.

Semrush's data makes the gap visible. A brand can appear frequently in AI answers and still receive very little referral traffic. The AI may have answered the question fully enough that the user has no reason to click. It may also mention the brand without linking to its website, or send the buyer towards a retailer, review site or marketplace instead.

This is why an improvement in AI visibility does not automatically produce a matching movement in GA4.

Where influence breaks down

The practical value lies in identifying the first point where performance drops. That tells a business which problem to solve first, instead of treating every weak result as an AI visibility issue.

Acting on the wrong layer can make a metric improve without changing the commercial outcome. More citations will not repair a weak landing experience, and better attribution will not make a brand more likely to be recommended.

The part the tools still don't answer

Taken together, these tools can show whether a brand is mentioned, cited, visited or credited with revenue. What they still struggle to explain is where the chain is breaking for a specific business.

Low referral traffic could mean the AI answer removed the need to click, the link went elsewhere, or the visit was hidden in Direct. Strong referral traffic with weak conversion points to a different problem entirely.

The missing layer is diagnosis: identifying which constraint is holding performance back, and therefore what the business should fix next.

Want to see where AI influence is breaking down for your business?

Run the AI Visibility Scan to check how AI engines currently describe and recommend your brand, or visit our services page to see how Xyra turns visibility, citation and conversion signals into a structured audit and action plan.

---

Frequently asked questions

Why are companies measuring AI search in such different ways?

There is no shared standard for connecting AI visibility with commercial impact. Peec, Semrush, AirOps and Profound each concentrate on a different part of the journey: referral traffic, mentions and citations, revenue attribution, or brand representation.

Is AI referral traffic accurately tracked in GA4?

Not fully. Clicks from Google AI Overviews appear as ordinary Google Organic Search, while visits without a usable referrer can appear as Direct. GA4 can capture identifiable AI referrals, but it cannot show every journey in which AI influenced the decision.

Does being mentioned by AI mean a brand will receive website traffic?

Not reliably. Semrush found that only two of the top 15 manufacturing brands for AI mentions and citations also ranked among the top 15 for AI referral traffic. Visibility and referral traffic measure different behaviours.

What is the difference between AI visibility and AI attribution?

AI visibility measures whether and how a brand appears in AI-generated answers. AI attribution tries to connect that exposure with a business outcome, such as a visit, lead or sale.

How does Xyra approach AI measurement?

Xyra examines the full journey from representation and evidence through to traffic capture and conversion. The aim is to identify where influence breaks down for the individual business, because each constraint requires a different fix. Learn more about the AI Readiness Audit →