Playbook · methodology
Attribute AI search revenue
Many buyers who find you through AI never click a trackable link: they ask, read and later type your name. This playbook combines self-reported attribution with orders and deals, so you can put a revenue figure next to your AI visibility.
- Goal
- A defensible share of revenue from AI search
- For
- E-commerce, SaaS, B2B lead generation
- Baseline
- 2–4 weeks of survey responses
- Re-measure
- Monthly, in rolling 3-month windows
Goal
Answer one question with data: how much revenue comes from people who found you through AI answers? The method pairs a “How did you hear about us?” question with your conversions and adds AI referral traffic from your analytics as a second, click-based signal.
Who it's for
- E-commerce shops on Shopify, WooCommerce or Shopware
- SaaS companies billing through Stripe
- B2B teams with leads in HubSpot, Salesforce, Pipedrive or another CRM
- Marketing leads who need to justify a GEO budget
Prompt set
Attribution is measured with a survey, not with prompts. Tracking purchase-intent prompts alongside it shows whether visibility and revenue move together. Ten templates.
Import it in AutoSEO under AI Visibility → Tracker → Import CSV (paste the text or upload the file).
Replace these placeholders before you import — the importer takes every line literally:
[category]— your product category[brand]— your brand name[competitor]— your strongest competitor[audience]and[use case]— a buyer segment and a job to be done
prompt,tags"Where can I buy a good [category]?","Purchase intent""Which [category] should I buy for [use case]?","Purchase intent""Best [category] for [audience] with good support","Purchase intent""Which [category] offers the best value for money?","Purchase intent""Where do people buy [category] online?","Purchase intent|Retail""[brand] or [competitor]: which should I buy?","Purchase intent|Head-to-head""Is [brand] worth the money?","Purchase intent|Brand""Is [brand] legit?","Purchase intent|Brand""What do customers say about [brand]?","Purchase intent|Brand""Does [brand] offer a free trial or returns?","Purchase intent|Brand"Engines to track
Track the engines your survey offers as answers, so you can compare visibility and self-reported revenue per engine.
| Engine | How it fits the survey |
|---|---|
| ChatGPT | Track it by default and offer “ChatGPT” as its own answer option. |
| Perplexity, Gemini, Claude, Microsoft Copilot | Offer them as separate options, so answers map to tracked engines. |
| Google AI Overviews and AI Mode | Respondents may just answer “Google”. Offer “Google AI answer” as its own option to separate it from classic search. |
Baseline measurement in AutoSEO
- 1
Run the seven-step setup
Attribution guides you through what to track, your platform, your forms, the survey, installing the snippet, conversions and live verification.
- 2
Ask the question
Ask “How did you hear about us?” at checkout, in your signup form or in your lead forms, with AI assistants as explicit options next to search, social and referral.
- 3
Connect conversions
Send purchases and deals with their value: through the website snippet, Stripe, Shopify, WooCommerce, Shopware or a CRM webhook for HubSpot, Salesforce, Pipedrive and others.
- 4
Verify, then wait
The last setup step confirms the snippet, the first response and the first conversion. Then collect responses for two to four weeks before you read the numbers.
Actions
- 01
Map answers to channels
Map free-text answers to channels, so “chatgpt”, “Chat GPT” and “the AI” all count as AI search.
- 02
Add AI referral traffic
Connect Google Analytics, Matomo or Piwik PRO. Analytics → Human Traffic shows visits referred by AI platforms — the clicks the survey can't see.
- 03
Compare channels by revenue
Put AI search next to organic search, paid, social and referrals by revenue and deal value, not only by number of responses.
- 04
Relate visibility to revenue
Compare revenue from AI search with the mention rate for the purchase-intent prompts over the same months.
- 05
Report it monthly
Build a monthly report with attribution next to visibility and export it as PPTX or PDF for stakeholders.
- 06
Refine the question between periods
If many answers are vague (“internet”, “Google”), refine the options — between measurement periods, never in the middle of one.
How to measure change
- Share of responses naming an AI assistant, per month
- Revenue and deals attributed to AI search, next to the other channels
- Survey response rate — low rates make shares unreliable
- Sessions referred by AI platforms in analytics, as a separate signal
- Mention rate for the purchase-intent prompts over the same period
Report a range, not a point
Self-reported data is honest but imprecise. Report AI search as a range across months and methods — survey and referral traffic — rather than as one exact figure.
Expected variance
Attribution numbers depend on sample size. With few responses per month, one or two answers can move the AI share noticeably. Wait for a meaningful number of responses before drawing conclusions, and look at rolling three-month windows.
Self-reported answers also depend on memory: people tend to remember the last touchpoint, not always the first. Treat them as one perspective next to click-based analytics.
Limitations
- People who don't answer the survey are missing from the data. If AI users answer more or less often than others, the shares are skewed.
- Survey answers are self-reported and can't be verified per person.
- Referral traffic misses buyers who read an AI answer and visit later directly.
- Visibility and revenue moving together is not proof of causation: seasonality and campaigns move both.
Questions about this playbook
Why not just use web analytics?
Analytics only sees visits with a referrer. Many people read an AI answer and come back later by typing your name, which analytics records as direct traffic or branded search. The survey closes that gap, and analytics adds the click-based view.
Where should I ask the question?
Where people convert: at checkout or on the order confirmation page, in the signup form, or in lead and demo forms. AutoSEO works with its own snippet and with form and survey tools such as Typeform, Tally, Jotform, Fairing, KnoCommerce and Zigpoll.
Which answer options should I offer?
Name the AI assistants individually (ChatGPT, Perplexity, Gemini, Claude, Copilot), add “Google AI answer”, classic search, social media, a recommendation, podcast or press, and a free-text “Other”. Randomize the order if your form tool supports it.
How many responses do I need?
There is no universal threshold, but shares based on a few dozen answers per month are rough. Use rolling three-month windows until you have enough volume, and always report the number of responses next to the share.
Can I import survey answers I collected before?
Yes. Attribution accepts CSV imports, so you can bring in answers from before you connected AutoSEO and compare periods.
Run this playbook in AutoSEO
Every feature is included in the free self-hosted edition and in AutoSEO Cloud.