THE AGENT NETWORK · SIX SURFACES
Six agents answer the buying question.
One integration reaches them all.
Each surface decides what to recommend from a different substrate - one reads your merchant feed, one reads public citations, one will not see you at all without a live MCP connection. Connect once and we handle the difference.
How each surface decides
The same catalog, read six different ways
Optimising for one surface does not carry to the next. This is the practical difference between them.
| Surface | Grounds its answer on | Your way in |
|---|---|---|
| ChatGPT | Citations + structured product data | Feed quality, citable pages, in-chat checkout |
| Google Gemini | Shopping Graph + merchant feed + schema | Feed completeness, Product markup |
| Perplexity | Live web sources, publicly cited | Specific, current, question-shaped pages |
| Amazon Rufus | Listing data, reviews, buyer Q&A | Attribute depth, review coverage |
| Microsoft Copilot | Bing index + Microsoft merchant data | Bing coverage, IndexNow, merchant feed |
| Claude | MCP tool calls, not a crawled index | A live MCP server on your catalog |
ChatGPT
ChatGPT shopping recommends products inside the chat based on structured product data and citations, then completes the purchase in-conversation.
Citations + structured product data
22% → 56%
House of Zelena's ChatGPT mention share over 6 months, moving #6 → #1 across 3 LLMs (Tru Commerce case study)
Google Gemini
Gemini synthesises product answers from Google's Shopping Graph, your merchant feed, and the structured markup on your pages.
Shopping Graph + merchant feed + schema
+89%
AI Share of Voice lift for ITC MasterChef in 30 days, category rank #4 → #3 (Tru Commerce case study)
Perplexity
Perplexity answers buying questions with an explicit citation list, then increasingly supports purchase inside the answer.
Live web sources, publicly cited
22% → 56%
House of Zelena's mention share across three LLMs over 6 months, moving #6 → #1 (Tru Commerce case study)
Amazon Rufus
Amazon Rufus answers product questions inside the Amazon app and site by reading structured product data, reviews, and buyer Q&A - not ad spend.
Listing data, reviews, buyer Q&A
+0.98%
AI Share of Voice lift for Amazon India across 6 categories in 6 weeks - measured first-party (Tru Commerce case study)
Microsoft Copilot
Copilot grounds its shopping answers in the Bing index, Microsoft Merchant Center data, and structured content on your site - then increasingly supports purchase inside the assistant.
Bing index + Microsoft merchant data
+0.98%
AI Share of Voice lift for Amazon India across 6 categories in 6 weeks, measured first-party (Tru Commerce case study)
Claude
Claude's commerce path runs through the Model Context Protocol rather than a crawled index.
MCP tool calls, not a crawled index
7×
MyMuse's real AI-driven revenue versus what GA4 reported - ₹81.2K/mo actual against ₹11.5K/mo attributed (Tru Commerce case study)