Introduction
Your audience is not a "static profile"; it's measurable behavior. With AI Overviews and agents like ChatGPT/Perplexity, new entry points and channels emerge. This playbook breaks down how to: define tasks, tag sources by agent,measure in-chat events, elevate your entity and control bots.
For whom: B2B/SEO/growth marketers who want to measure real intent and attribution from AI in 90 days.
TL;DR — actionables
- Define 3–5 user tasks and map conversational questions.
- Tag links with agent UTMs (
utm_source=chatgpt|perplexity|gemini). - Create GA4 in-chat events:
generate_lead_in_chat,schedule_meeting_in_chat. - Publish schema (Organization/Service/ContactPoint) and entity hubs to elevate ECS.
- Control bots with robots.txt: Google-Extended/PerplexityBot/GPTBot by paths.
- Seal images with Content Credentials (C2PA) for trust and citability.
Quick answer
Identify audience by instrumenting intent: tasks → agent UTMs → in-chat events → solid entity (schema/hubs) → bot control. It returns who arrives, from where and with what intent.
Step-by-step development
1) Technical control and tagging (B2B)
robots.txt by bot
Rules for crawlers (GPTBot, Google-Extended, PerplexityBot, etc.). Controls discovery and AI usage; allows exceptions by paths.
Where it goes: Real file at /robots.txt (site root)
How to verify: Open /robots.txt (200 OK) and check logs/CDN. Purge cache if you change.
UTMs by agent
Parameters for attribution by model/assistant (chatgpt, perplexity, gemini). Allows segmenting AI traffic in GA4/CRM.
Example: ?utm_source=chatgpt&utm_medium=ai&utm_campaign=q1_2025
Where it goes: Links/CTAs; not in the body as plain text.
How to verify: Test links and check GA4 by source/medium/campaign
2) Editorial and brand operations
- Entity hubs with stable @id (docs/entities, help, research).
- Copyable answers to induce citations in AIO/Chats.
- Content Credentials in images for traceability.
3) Measurement and experiments
- Unique coupons by channel and cohorts by UTM.
- Compare MQL/SQL/Demo rate with custom events.
Limitations and operational reality 2025
- Unequal bot compliance;
llms.txtis experimental. - Partial attribution; use cohorts and UTMs.
- AI errors/hallucinations; reinforce official pages and credentials.
AEO/GEO/visibility metrics (B2B table)
| Metric | Definition | How to measure |
|---|---|---|
| SGV | Visibility generated by search/AI for your brand/services | Query/mention monitoring + cited links in AIO |
| Attribution Rate from AI | (AIO with link to your site) / (Total AIO in your topic) | Biweekly sampling + UTMs by agent |
| ECS (Entity Consistency Score) | Entity consistency (name, @id, schema, hubs) | Schema/hub audit + biweekly diff |
| Time-to-Answer | Time to deliver clear answer | Benchmark /help + engagement |
| MQL from AI | MQL with first touch utm_source=chatgpt|perplexity|gemini | GA4 cohorts + CRM |
| SQL/Demos in-chat | % MQL→SQL with generate_lead_in_chat/schedule_meeting_in_chat events | GA4 + CRM |
