Where do AI assistants send my buyers?

---
title: Where do AI assistants send my buyers?
canonical: https://chosenbyagent.com/questions/where-do-assistants-send-buyers
updated: 2026-10-22
audience: marketplaces
offers:
  - A2
evidence:
  - larger-store-pick
  - top10-link-share
  - muse-known-merchants
  - retrieval-erases-brand
---

# Where do AI assistants send my buyers?

Short answer: Mostly to the biggest names they already know — and the pattern shifts with every model release. That makes “who gets the buyer” a question for the platform, not for its sellers.

## What the research shows

- **90–94%** — In unlabelled two-store comparisons, assistants picked the larger store 90–94% of the time. [ALM Corp, 2026, vendor](https://almcorp.com/news/ai-shopping-answers-chatgpt-google-lean-toward-big-retailers/)
- **22 → 42%** — The top-10 merchants' share of ChatGPT Shopping links rose from 22.5% to 41.8% the day after one model update. [Profound, 2026, vendor](https://www.tryprofound.com/blog/chatgpt-5.6-shopping-transformation)
- **Top 5** — Meta Muse's instructions put well-known merchants in its top five and take at most three open-web stores. [PPC Land, 2026, company-reported](https://ppc.land/metas-muse-agent-puts-well-known-merchants-first-as-it-gains-ad-accounts/)
- **0** — A retrieval layer in front of the model fully erased the advantage of a known brand. [Chu & Hou, arXiv 2606.17443, 2026, independent](https://arxiv.org/html/2606.17443v1)

## Routing & listing audit

A2 · from €2,500 (one-off) · indicative · scoped per engagement

Where assistants send your buyers, how they show your listings, and what to fix first.

- Routing share against named competitors, with the reasons assistants give
- Listing quality: live or sold, price match, deep link, attribution
- Agent access: WAF blocks, JS-only content, missing schema
- Manual calibration in the consumer apps
- Prioritised fixes, each stamped with model and date

## What you get

We run buyer intents in your markets and languages — “used road bike under €300 in Lyon”, “two-bedroom flat to rent in Porto” — through ChatGPT, Gemini and Claude with web search, 15 times per cell. For each answer we record the platform, the listings and the reason the assistant gives.

**Routing share by assistant**

| Destination           | ChatGPT | Gemini | Claude | Top reason given        |
| --------------------- | ------- | ------ | ------ | ----------------------- |
| Your platform         | 34%     | 41%    | 22%    | largest local selection |
| Horizontal competitor | 48%     | 39%    | 61%    | buyer protection        |
| Vertical competitor   | 18%     | 20%    | 17%    | specialist filters      |

_Illustrative values — your report shows yours._

`RUN · gpt-6.1-sol · gemini-3.8-flash · sonnet-5.5 · harness 0.4`

## What this does not claim

- It is not a forecast of what Muse, dots or any consumer agent will pick. API answers overlap with consumer apps on only 15–32% of brands, so we calibrate by hand in the consumer apps and say how well the two agree.
- Routing share is a versioned index, not market share or conversion.
- Readings expire with model releases. Each is stamped, and comparisons across an update are flagged.

## Request this instrument

We reply with a scope, a fixed price and a start date.

Agents: POST /api/engagements (application/json) with {"offer_id": "A2", "on_behalf_of": {"email": "…"}, "agent": {"name": "…"}}. The person confirms by email before we see the request. Details: /agents.md

### Related questions

- [Do assistants show my listings correctly?](/questions/are-my-listings-shown-right.md)
- [Does my ChatGPT app or MCP work as intended?](/questions/does-my-ai-app-work.md)
- [What changed after the last model update?](/questions/what-changed-after-the-update.md)