2026-09-17, Copilot · Rank Tracking · GEO

Bing rank tracking and Copilot visibility: what to use

Bing visibility now has two separate parts. Bing web rankings are positions of your pages on Bing’s results page. Copilot visibility is whether Microsoft’s assistant mentions and cites you when it answers. They need different data sources. For Bing web rankings, use Bing Webmaster Tools data for your own site, a rank tracker that supports Bing, or a SERP API with a Bing engine. For Copilot, use an answer monitoring API. This API covers Copilot; it does not return Bing web results, and there is no Bing endpoint in its OpenAPI document.

What changed for Bing data

Two Microsoft changes shape the options, both checked 2026-09-17:

• The Bing Search APIs are retired. Microsoft retired them on August 11, 2025 and decommissioned existing instances (lifecycle notice).
• The suggested replacement does not return results. Grounding with Bing Search, now documented under Microsoft Foundry Agent Service, gives an agent’s model web grounding, but the documentation states that developers and end users do not have access to the raw content it returns (Microsoft docs). You get the model’s answer with citations, not a ranked list.

So there is no first-party API that returns Bing’s ranked results for arbitrary queries.

Options for Bing web rankings

Bing Webmaster Tools data

For your own verified sites, the Bing Webmaster API is the first-party source. Its GetQueryStats method returns per query: Query, Impressions, Clicks, AvgImpressionPosition, AvgClickPosition and Date, and Microsoft’s reference says the data is updated every week (GetQueryStats). Microsoft’s documentation also says the legacy SOAP and POX interfaces retire on August 31, 2026, so new integrations should use the REST interface.

Average position from Webmaster data is an aggregate over your impressions, which makes it good for trends on your own site and useless for competitors or for queries where you have no impressions.

Rank trackers that list Bing

From vendor pages checked 2026-09-17:

We found no Bing rank tracking on Ahrefs’ or Wincher’s pages. Entry prices for these tools are compared in rank tracking software vs building on an API.

SERP APIs with a Bing engine

Several SERP APIs include Bing; the vendors and their surfaces are compared in best SERP APIs. If you need Bing web positions as data for any query, that is the category to evaluate. This API is not in it.

Copilot visibility with this API

Copilot visibility is how often, and how prominently, Copilot mentions your brand or cites your pages when answering the prompts your customers ask.

On grounding: Microsoft’s documentation for Microsoft 365 Copilot says Copilot may fetch information from the Bing search service, turning a prompt into a short search query sent to Bing (Microsoft Learn). The transparency note for consumer Copilot says it is grounded in web search results (Microsoft). Either way, the searches Copilot runs and the pages it cites are what you can observe, and both are in the response.

The request

POST /v1/monitor/copilot takes:

curl -X POST https://api.answerline.dev/v1/monitor/copilot \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "prompt": "What is the best project management tool for a 10-person agency?", "country": "US" }'

The response

result contains:

• text: the answer.
• sources[]: sources referenced, with position, url, label and description.
• citationPills[]: the inline citation chips, with citationPillId, url, domain and position.
• markdown, rawResponse and rawWire when requested.

The web searches Copilot ran, product cards and map results are coming; they are not returned yet.

Copilot responses have no entities[] field, so brand mentions come from matching names in text. Monitoring Copilot with an API walks through every field.

Turning Copilot answers into rankings

Define four numbers per prompt, market and run:

1. Mention: your brand name, or an alias, appears in text.
2. Citation: a sources[].url or citationPills[].domain is on your domain.
3. Citation rank: the lowest position among your sources, empty when not cited.
4. Fan-out overlap: how many searchQueries correspond to keywords you track on Bing.

import re
from urllib.parse import urlparse

BRAND = re.compile(r"\b(acme|acme pm)\b", re.I)
MY_DOMAIN = "acme.example"

def host(url: str) -> str:
    h = urlparse(url).netloc.lower()
    return h[4:] if h.startswith("www.") else h

def copilot_row(prompt_id: str, market: str, result: dict, tracked_keywords: set[str]) -> dict:
    mine = [s["position"] for s in result.get("sources", []) if host(s["url"]).endswith(MY_DOMAIN)]
    pills = {p["citationPillId"] for p in result.get("citationPills", []) if p.get("domain", "").endswith(MY_DOMAIN)}
    queries = [q.lower() for q in result.get("searchQueries", [])]
    return {
        "prompt": prompt_id,
        "market": market,
        "mentioned": bool(BRAND.search(result.get("text", ""))),
        "cited": bool(mine),
        "citation_rank": min(mine) if mine else None,
        "inline_pills": len(pills),
        "search_queries": queries,
        "fanout_tracked": sum(1 for q in queries if q in tracked_keywords),
    }

Connecting Copilot to Bing keywords

searchQueries is the bridge between the two halves. Collect them over a few weeks and:

1. Count how often each query appears across your prompt set.
2. Compare frequent fan-out queries with the queries in your Bing Webmaster GetQueryStats export.
3. Queries Copilot runs often, where your site has no impressions in Webmaster data, are content gaps for both Bing and Copilot.
4. Add frequent fan-out queries to your Bing rank tracker’s keyword list.

Query fan-out covers the concept. ChatGPT’s search-query list comes back empty on the answers served today, so this technique needs an engine that returns its searches.

Scheduling a Copilot program

Copilot answers vary between runs, so treat each run as a sample and aggregate over several. Submit prompts as async tasks in batches of up to 500:

import os
from datetime import date

import requests

API = "https://api.answerline.dev"
HEADERS = {"Authorization": f"Bearer {os.environ['ANSWERLINE_API_KEY']}"}

prompts = {
    "pm-agency": "What is the best project management tool for a 10-person agency?",
    "pm-vs": "Acme PM vs Trello for client work",
}
run = date.today().isoformat()

tasks = [
    {
        "taskType": "COPILOT",
        "payload": {"prompt": text, "country": "US"},
        "idempotencyKey": f"copilot:{run}:{pid}:US",
        "webhook": {"url": "https://example.com/hooks/copilot"},
    }
    for pid, text in prompts.items()
]
resp = requests.post(f"{API}/v1/async/task/batch", headers=HEADERS, json=tasks, timeout=60)
resp.raise_for_status()
print(resp.json()["summary"])

Results arrive at the webhook as the task status body; verify Webhook-Signature before storing (how). Copilot takes any country; a US state is refused with a 400.

What it costs

A Copilot request is $0.00274 on every tier. Copilot’s include flags add nothing to the price.

New accounts get 100 free requests, which cover 100 Copilot tasks. Tier prices are on /pricing.

A combined Bing and Copilot report

Pitfalls

• Expecting Bing results from this API. It has no Bing web endpoint; use Webmaster data, a tracker or a Bing SERP API.
• Treating Grounding with Bing Search as a data source. It does not return raw results.
• Comparing Webmaster average positions with tracker positions. One is an impression-weighted average for your site, the other a point observation.
• Single-run Copilot conclusions. Sample repeatedly and report rates.
• Matching brands only on domains. Copilot can mention you without citing you; check text too.

Start with the Copilot engine page or the quickstart.

Keep reading