How to do competitor analysis with Microsoft Copilot

Why use Microsoft Copilot for competitor analysis

Microsoft Copilot can produce a solid first-pass competitor analysis if you give it a fixed structure, turn on web search, and make it put a source next to every claim. Its advantage is where it sits. It sits next to Word, Excel, Teams, and Outlook, so many people already have it at work, and on the paid work license it can read your company's own files as well as the web. That means you can combine the internal notes you already have with what Copilot finds on a competitor's site, in one place. This guide gives you the exact prompt, the method around it, a worked example, and a checklist for grading what comes back.

One thing before we start: we build Competely, a competitor analysis tool, so this exact job is what we work on all day. Everything below runs entirely in Copilot. At the end we will tell you where a dedicated tool does the job better, and where it honestly does not matter.

A quick note on names, because Microsoft uses "Copilot" for several products. This guide is about the Microsoft Copilot assistant for work and personal use. It is not about GitHub Copilot, which is a coding assistant licensed separately and covered in our AI agents guide. Within the assistant, three tiers matter here. The free consumer Copilot (copilot.microsoft.com, with a personal Microsoft account) answers from web search but has no access to work data. Microsoft Copilot Chat is included with eligible Microsoft 365 business subscriptions when you sign in with a work account. It is grounded in the web and reads files you upload. The paid Microsoft 365 Copilot add-on adds your work data through Microsoft Graph, Copilot inside Word, Excel, and PowerPoint, and agents such as Researcher. Microsoft is also renaming Microsoft 365 Copilot to Microsoft Copilot, so both names appear for now. The prompt below works on all three tiers, and features that depend on the paid license or on admin settings are called out where they come up.

What Copilot gets right, and where it hits a ceiling

Copilot's strength for this job is grounding. When web search is on, Copilot turns your prompt into a few Bing queries, reads the results, and shows linked citations under the answer. On the paid license it also pulls in relevant emails, files, and meetings you already have access to, so what your sales calls said about a competitor can sit next to what its pricing page says today. Researcher, the deep research agent on the paid license and on Microsoft 365 Premium, goes further: it asks clarifying questions, runs a multi-step investigation across the web and your work files, and returns a structured, source-cited report. Microsoft's own small-business tips recommend asking Copilot who your competitors are and then for a chart of their strengths, which is what it does well out of the box.

The ceilings are specific. Citations are not universal: Copilot Chat shows web citations, but Microsoft's documentation says they are not available in the Copilot pane inside Word or PowerPoint. The prompt works around this by asking for the URL inside each sentence. Web search can also be off without you knowing: an admin can disable it for the whole organization, and each user has a "Web content" toggle. When it is off, Copilot answers from training data and your files only, and competitor pricing goes stale. Consistency between runs is weak: the same question on two competitors returns a table for one and bullets for the other. Pinning the headings fixes most of that. And Copilot does not watch competitors for you. Scheduled prompts exist on the paid license (more below), but they re-run a prompt, they do not tell you what changed.

Two practical limits before you start. Researcher runs are slow by design: Microsoft's FAQ says simple queries take under five minutes, complex ones 10 to 45 minutes, and it lists a cap of 25 Researcher queries per user per month. File uploads are on by default, but Microsoft only says that "file upload limits apply" without publishing numbers, and an admin can turn uploads off.

The method, step by step

1. Frame the eight dimensions before you type

"Tell me about Acme Analytics" produces a wandering essay. The prompt above hands Copilot eight labeled dimensions (Marketing & Messaging, Product & Features, Pricing & Sales, Audience & Customers, Customer Sentiment, Company Profile, Market & Opportunity, SWOT Analysis) and insists on the headings word for word. Same structure in, comparable output out. These are the same eight dimensions a Competely report uses, so if you later compare your output with a finished analysis, the sections line up. Fill in the two bracketed lines before you send it.

2. Run it with web search on, or in Researcher

In Copilot Chat, check that the "Web content" toggle is on before you send the prompt. If it is grayed out, your admin has disabled web search, and nothing in the prompt can turn it back on. If you have the paid license or Microsoft 365 Premium, open Researcher instead (under Agents, or under Tools in Copilot Chat), paste the same prompt, and answer its clarifying questions or tell it to go ahead. Researcher has its own web search toggle in the input box, so confirm that one too. Expect to wait, because a full eight-section report is the kind of complex query Microsoft says can take 10 to 45 minutes. Given the monthly cap, use ordinary chat for small follow-up questions.

3. One competitor per chat

Give each competitor its own chat or Researcher session. When two competitors share a conversation, details blur and a price from one attaches itself to the other. Name the chat after the competitor and the date so you can find it again, and use the same mode (for example "Think deeper") for every competitor so the depth is comparable.

4. Keep the results in a Page or a Word document with the same headings

Chat history is a poor filing system. Copilot Pages turn a response into an editable page you can share with colleagues, and Microsoft says they are available to work accounts with SharePoint or OneDrive storage, even without a Copilot license. Select the page button under the response and the analysis opens beside the chat. From a Page you can convert to a Word document through the Create menu, although Microsoft notes this is rolling out gradually. Whatever you use, keep one document per competitor with the same eight headings in the same order, so you can put them side by side later. A Copilot Notebook can hold the finished set, but Copilot does not search the web inside a notebook, so use it for comparing saved analyses rather than gathering new facts.

5. Verify pricing on the live page

Open the competitor's pricing page in another tab and compare it line by line with what Copilot returned: price, billing period, seat minimums, usage limits, anything marked "starts at." The live page wins every disagreement. Do the same for anything described as recent: a launch, a funding round, a rebrand. Open the source Copilot cited and check the date.

A worked example: Acme Analytics

Acme Analytics is a fictional competitor, and the table below is an illustration of a good Copilot run. The point is the shape: a claim, a source in the same sentence, and an honest gap where no source exists.

Two things mark a run you can trust: "no public source found" where the fact was unavailable, instead of a plausible funding number, and a research date with pricing and the "coming this fall" feature flagged as the claims most likely to go stale first.

Ten checks before you trust the output

Copilot will return something that looks finished. These checks tell you whether it is, and each grades something the prompt asked for.

1. Prices match the competitor's live pricing page, plan by plan: price, billing period, seat minimums, usage limits.
2. Every factual claim carries a source URL in its own sentence, and you have opened the sources behind anything you plan to act on.
3. Every "no public source found" line is resolved: you found the fact elsewhere, or you accepted it as unknown.
4. Messaging quotes match the competitor's site word for word rather than a paraphrase.
5. Features are split into live today versus announced, and nothing from a roadmap post is listed as shipped.
6. Recent moves each carry a date you can check.
7. Company numbers (funding, team size, customers) each carry a dated source.
8. Sentiment quotes link to reviews you can open, not to summaries of reviews.
9. Every competitor came back under the identical eight headings, so the analyses line up side by side.
10. The research date is on the document, and if the run used work files as well as the web, you know which claims came from which.

Start with pricing and the recent items. Those change fastest, and they are usually what you act on.

Keeping the analysis current

Everything above produces a snapshot, and competitors keep moving after you file it. You have three honest options.

Do it by hand. Set a recurring calendar reminder, re-run the same prompt in a new chat, and compare the new document with the old one. This is free and it works as long as you actually do it.

Build it in Copilot. If you have the paid Microsoft 365 Copilot license, Copilot has scheduled prompts. In Copilot Chat, Teams, or Outlook you hover over a prompt, choose "Schedule this prompt," set when and how many times it runs, and optionally get an email when the response is ready. You can have up to 10 scheduled prompts, and each schedule runs a limited number of times before it expires (a third-party write-up from February 2026 put the cap at 15 runs). Each run lands in your Chats list as a fresh answer, so there is no change log: you still compare the new one with the old one yourself. Scheduled prompts also depend on your admin leaving connected experiences enabled, and they are not available without the paid license.

Buy it. A monitoring product does the re-checking on a schedule and sends you what changed. You trade flexibility for not maintaining anything. That option is where our own product comes in.

Where Competely fits, and where it does not

Competely automates the method above. You enter your product and your competitors, and it analyzes each one across the same eight dimensions the prompt uses, more than 100 data points in total, each with its own source link. An analysis takes about 30 minutes, and because the structure never varies, the analyses line up side by side without cleanup. Monitoring is included on every paid plan: on a schedule you set (every two weeks or monthly on any plan, weekly on Pro and Scale) it re-reads each competitor's public web pages, flags what changed, and emails you a short brief. You can read the full Figma vs Adobe XD sample analysis without an account.

Now the limits. Competely has no Microsoft 365 integration. It does not read your SharePoint, Teams chats, or Outlook, and it does not appear inside Word or Copilot. If what you like about Copilot is that it combines internal notes with the web, Competely does not do that half. It watches public web pages only: no social listening, no press tracking, no win-loss interviews, no CRM integration, no API, and no SOC 2 report. Exports are CSV and PDF, and sharing is a link anyone can open. There is no free trial of the full product, though a free account with no credit card gets you a real but partial analysis against two competitors, with depth, exports, and monitoring on the paid plans. It is self-serve with no sales call, from $39/mo, billed monthly or yearly, cancel anytime. If Copilot plus a calendar reminder covers you, you do not need it. What it sells is a fixed structure and never doing the re-checking yourself.