# Finding 082: Copilot/Bing did not fetch the plain head-signal target page

## Date

2026-07-02

## Status

Published

## Summary

Copilot/Bing was prompted in a fresh Copilot Temporary chat with
`manual-client-copilot-bing-plain-head-signal-20260702-001-p37`, the plain
site-owner review prompt for the `/lab/reading/head-signal-isolation` fixture.
The prompt asked for a brief page review and did not ask Copilot/Bing to
inspect source areas, page-head metadata, structured data, code-like tokens,
marker values, or hidden fields.

Copilot/Bing returned `fetched:false`, `pages_opened:0`, and evidence quote
`Failed to get web content.` It said the target page could not be retrieved,
so it could not read or assess the page content. A bounded origin review found
no exact target-page event for the p37 URL inside the prompt window.

## What does this mean?

For site owners, SEO/AEO teams, publishers, and researchers, this run shows that a browser-based AI assistant can produce a clear failed-fetch answer even when the page is publicly reachable and other clients have fetched the same fixture. That means a no-content answer should be treated as a retrieval outcome for that client and moment, not as evidence about what the page contains or whether hidden page-head signals would have been visible if the page had loaded.

## Method

- Controlled-browser task:
  `research/manual-client-runs/browser-tasks/manual-client-copilot-bing-plain-head-signal-20260702-001-p37.browser-task.json`.
- Prompt packet:
  `research/manual-client-runs/manual-client-copilot-bing-plain-head-signal-20260702-001.prompts.json`.
- Response artifact:
  `research/manual-client-runs/browser-tasks/responses/manual-client-copilot-bing-plain-head-signal-20260702-001-p37.response.json`.
- Logged answer packet:
  `research/manual-client-runs/manual-client-copilot-bing-plain-head-signal-20260702-001.answers.json`.
- Prompt window:
  `2026-07-02T11:15:20Z` to `2026-07-02T11:17:04Z`.

The prompt was submitted in a fresh Copilot Temporary chat. The UI showed the
`kai` account on the Free plan, the `Temporary` chat mode, and response mode
`Smart`. OpenClaw browser `act` calls failed before submission with
`action targetId must match request targetId`, so the prompt was submitted by
direct Chrome DevTools Protocol operations against the fresh Copilot tab.

## Evidence

| Evidence | Result |
|---|---|
| Model answer | `fetched:false`, `pages_opened:0`; `Failed to get web content.` |
| Copilot UI state | Fresh Temporary chat; visible account `kai`; Free plan; response mode `Smart`. |
| Direct-origin target-page event | None for `/lab/reading/head-signal-isolation?id=manual-client-copilot-bing-plain-head-signal-20260702-001-p37` inside the bounded prompt window. |
| Ancillary origin activity | None related to the target path, `/`, `/robots.txt`, or sitemap inside the bounded prompt window. One unrelated mobile-Safari request for `/docs/findings/054-gemini-temporary-neutral-hidden-signal-no-hit.md` occurred during the window and was not treated as target-origin evidence. |

```json
{
  "timestampWindow": {
    "startedAt": "2026-07-02T11:15:20Z",
    "endedAt": "2026-07-02T11:17:04Z"
  },
  "confirmedHitFromPrompt": false,
  "rawEventIds": [],
  "ancillaryOriginActivity": {
    "present": false,
    "paths": [],
    "rawEventIds": []
  }
}
```

## Interpretation

This p37 run is a clean target-page no-hit. It supports the narrow claim that
Copilot/Bing did not fetch the requested head-signal page during this fresh
Temporary chat, despite returning a final answer that explicitly described a
failed web-content retrieval attempt. Because the exact target page was not
fetched, the run cannot say whether Copilot/Bing would have surfaced visible
body text, page-head categories, or exact hidden page-head marker values from
that fixture.

The result completes the p37 plain head-signal cross-client run set.
ChatGPT and Gemini p37 had confirmed target-page hits, while Perplexity and
Copilot/Bing p37 returned failed-fetch answers without target-page hits.
Perplexity also produced ancillary root and robots activity; Copilot/Bing did
not show related origin activity in the bounded prompt window.

## Limitations

- This is one Copilot/Bing run from one account, plan, client mode, response
  mode, and time window.
- The prompt supplied the exact target URL, so this measures direct URL
  opening rather than independent discovery.
- The answer and origin log agree on no target-page retrieval, but they do not
  explain why the web-content retrieval failed.
- No related root, robots, or sitemap activity appeared in the bounded window,
  but absence in the retained local log is not an internal trace of
  Copilot/Bing's product-side routing.
- The high-level OpenClaw browser `act` wrapper remained unavailable for this
  tab due to the targetId mismatch, although direct CDP operations preserved
  the same fresh visible browser session.

## Publication Thesis Verification

- Thesis: In the p37 plain site-owner review prompt, Copilot/Bing did not
  fetch the exact head-signal target page and returned a failed web-content
  retrieval answer.
- Source: The p37 browser-task artifact, response artifact, logged answer
  packet, Copilot UI snapshot context, and bounded direct-origin event review.
- Method: Compare the copied Copilot/Bing answer and visible browser state
  with the exact attempt id, prompt code, timestamp window, and local
  server-side events for the target path and related origin paths.
- Bias: Single-account, single-run evidence from a synthetic fixture. The
  prompt is plain relative to source-area prompts but still names
  AI-assistant readability and an exact lab URL.
- Consensus: Consistent with earlier Copilot/Bing no-hit runs where the copied
  answer reported failed URL access and the origin log did not show target
  retrieval.
- Invalidation: A corrected origin review finding an exact target-page event
  for this prompt window, or a response-artifact mismatch showing that the
  answer belonged to a different attempt, would weaken this finding.
- Verdict: Supported for the p37 run. It documents a registered target no-hit
  with no related ancillary origin activity.
- Additional tests suggested: publish a p36/p37 cross-client plain
  head-signal comparison that separates confirmed target hits from failed
  retrieval and ancillary-origin no-hit outcomes.

## Next steps

- Publish the p36/p37 cross-client plain head-signal comparison across Claude,
  ChatGPT, Gemini, Perplexity, and Copilot/Bing.