Not all AI agents behave the same way. ChatGPT, Gemini and Perplexity have different recommendation behaviors, different data sources and different levels of product-level confidence. A product that converts well via one agent may barely surface on another. This playbook shows you how to read your channel data and make one clear decision about where to focus next.
Goal: You leave this session knowing which AI agent is your highest-value channel and what you should do differently because of it.
Time to complete: 20–30 minutes.
Features used: AgentIQ Overview → Agent Traffic Analytics · AgentIQ Overview → AI Agents Distribution · AgentIQ Overview → Best Attributed Products
Step 1: See Which AI Agents Are Sending You Traffic
AgentIQ Overview — Agent Traffic Analytics showing the AI Agents Distribution breakdown by channel
Go to AgentIQ Overview.
Scroll to the AI Agents Distribution section. This shows you the breakdown of sessions across ChatGPT, Claude, Gemini and other sources — both in raw session counts and as a percentage of your total agentic traffic.
Note which agent is your largest source of sessions. That's your most active recommendation channel.
Now note which agent's share is growing fastest, if any. Volume and trajectory are different signals — a smaller agent that's growing quickly may deserve more attention than a large one that's plateaued.
Step 2: Identify Products That Perform Differently by Agent
Product Leaderboard — showing products by appearance rate, optimization status and warning tags
Different AI agents have different recommendation behaviors. A product that converts well via ChatGPT may not surface the same way on Gemini. Finding these patterns tells you where your APL optimization is working — and where it isn't.
In AgentIQ Overview, scroll to the Best Attributed Products section.
Look at which products are driving the most attributed revenue. These are the products your best agent sources are recommending with enough confidence to close the sale.
Check the Product Leaderboard section. For any product with a high appearance rate but low optimization status, note the warning tags — these are the specific gaps that are limiting that product's performance across agents.
Click Fix Now on any high-appearance product with unresolved warnings. That combination — high visibility, low optimization — is where fixing the listing will have the largest immediate impact.
Step 3: Find Your Highest-Converting Agent Source
Session volume and conversion rate are two different things. An agent that sends you fewer sessions but converts them at twice the rate is worth more than one that sends high volume that bounces.
In the AI Agents Distribution section, compare conversion rates across channels.
Identify the agent with the highest conversion rate. That's your highest-quality traffic source — the one where shoppers are most ready to buy when they arrive.
Check whether that agent is also your largest volume source. If it is, you're in a strong position. If it isn't, you have a clear opportunity: increasing your visibility on your highest-converting agent will have a disproportionate revenue impact.
Step 4: Make One Clear Decision and Act on It
The purpose of this session is to leave with one clear decision — not a list of observations. Here are the three most common decisions that come out of this analysis and what to do next for each:
Decision A: Double down on your highest-converting channel. Your best-converting agent isn't your biggest volume source. Use Playbook 1 or Playbook 3 to improve your ranking on that specific agent's prompts — run target prompts in the Shopify Catalog and create Agentic Product Listings for the products that should be appearing but aren't.
Decision B: Fix the products with the highest appearance rate and unresolved warnings. You're showing up — but something in the listing is breaking conversion after the recommendation. Use Playbook 5 to resolve the gaps for your most-recommended products.
Decision C: Investigate a growing channel before it becomes large. An agent that's growing its share of your agentic traffic is worth optimizing for now — before competitors do. Build a prompt library using that agent's query style and create APLs for your best products against it.
Related Articles