ChatGPT Prompts for Competitor Analysis (Without Made-Up Facts)

Ask ChatGPT "who are my competitors?" and you'll get a confident list. Some names will be real, some will be outdated, and a few may not exist. OpenAI's own help center says ChatGPT can produce "fabricated quotes, studies, citations or references to non-existent sources" and may sound confident even when it's wrong (OpenAI Help: Does ChatGPT tell the truth?). That's a bad foundation for a pricing page or a pitch.

This guide is for solo founders who want a competitor analysis they can trust in an afternoon. You'll get three copy-paste prompts: one to list every real alternative your buyer has, one to turn pages and reviews you collected into an evidence-labeled grid, and one to find a positioning angle you can defend. ChatGPT does the sorting. You supply the facts.

What ChatGPT is good at here (and what it must not do)

A classic competitor analysis assesses the strengths and weaknesses of current and potential competitors, often with a simple "competitor array": define the market, list competitors, list what customers care about, then rate each competitor on those factors (Wikipedia: Competitor analysis). That structure is exactly the kind of work ChatGPT handles well.

What it should not do is supply the facts. Prices change, features ship, and pages go offline. Treat anything ChatGPT says about a competitor from memory as a lead to check, not evidence. If your plan includes web search, you can let it look things up, but still open every cited link yourself before you rely on it.

The rule for all three prompts: every claim gets a source or a label. FACT means you can point to the page. INFERENCE means it's a reasonable reading of the evidence. ASSUMPTION means it's a guess you still need to test.

What to gather first (about 45 minutes)

• Your one-line offer. Who it's for and the job it does, in plain words.
• 3–5 alternatives your buyers mention. Pull them from customer calls, support emails, and "we use X today" comments. If you haven't talked to buyers yet, start with the customer interviews guide.
• For each alternative: the homepage headline (copied word for word), the pricing page (copied, with today's date), and 5–10 real review excerpts or forum comments, each with its link.
• The non-product options. Spreadsheets, a freelancer, an agency, a template, or doing nothing. The U.S. Small Business Administration's guide to competitive analysis explicitly tells you to look at "indirect or secondary competitors who may impact your success" (SBA: Market research and competitive analysis).

Paste raw text, not summaries. Your summary already contains your bias.

Prompt 1: Map every real alternative (not just look-alike products)

Your real competition is whatever the buyer would do if you didn't exist. This prompt keeps ChatGPT from stopping at the obvious names.

You are helping me map the alternatives my buyer has today. Do not invent company names, prices, or features. If you suggest an alternative I didn't list, mark it SUGGESTION - VERIFY.

My offer (one line): [WHO IT'S FOR + THE JOB IT DOES]
Buyer: [ROLE / CONTEXT]
Alternatives buyers have mentioned to me: [LIST, WITH WHERE I HEARD EACH]
What buyers did before they found me (if known): [NOTES]

Output a table with these columns:
Alternative | Type (direct product / adjacent tool / service or freelancer / DIY workaround / do nothing) | When a buyer would pick it | What it costs them (money, time, or risk - only from my notes, otherwise "unknown") | Source (my note, or SUGGESTION - VERIFY)

Then:
1) Name the 3 alternatives I should research first and say why in one sentence each.
2) List 3 questions I should ask buyers to confirm this map.

Rules: "do nothing" must appear as a row. Do not rank alternatives by market share or popularity.

What good output looks like: a table where "do nothing" and the spreadsheet workaround sit next to the software, and every suggested name is flagged for checking. If ChatGPT ranks companies by "market share," delete that part. It has no reliable way to know.

Prompt 2: Turn pasted pages and reviews into an evidence grid

Now feed it the raw material you collected for your top three alternatives. One conversation per research session keeps the context clean.

Build a competitor grid using ONLY the material I paste below. Do not add facts from memory. If something isn't in the material, write "not in sources".

My offer (for context): [ONE LINE]
Buying criteria my customers say matter most (from interviews): [3-5 CRITERIA]

For each competitor, I pasted: homepage headline (verbatim, with date), pricing page text (with date), and review excerpts (with links).
---
[COMPETITOR A MATERIAL]
---
[COMPETITOR B MATERIAL]
---
[COMPETITOR C MATERIAL]
---

Output:
1) A grid: rows = competitors; columns = Who they target | Main promise (quote their headline) | Pricing model and listed price (with date) | Strength buyers praise | Complaint buyers repeat | How each scores on my buying criteria (Strong / OK / Weak / not in sources).
2) Under the grid, label each non-obvious claim as FACT (quote the source line), INFERENCE, or ASSUMPTION.
3) Complaints that appear in 2+ reviews, with the quotes side by side.
4) Where each competitor is honestly stronger than me.

Rules: quote, don't paraphrase, for headlines and complaints. Never invent a review, a price, or a number of customers.

Point 4 matters. A grid where you win every row isn't analysis, it's a sales slide, and buyers can tell. Add the date to the top of the grid. Competitor pages change, and an undated grid quietly goes stale.

Prompt 3: Find a positioning gap you can actually defend

In marketing, positioning is how buyers perceive a product compared with the alternatives they know (Wikipedia: Positioning (marketing)). Use the grid to find a specific buyer and situation where you're the obvious choice.

Using only the competitor grid and review complaints below, help me find a positioning angle I can defend. Do not invent customer numbers, results, or testimonials.

Competitor grid + repeated complaints:
---
[PASTE OUTPUT FROM PROMPT 2]
---
What my product actually does today (no roadmap items): [3-5 BULLETS]
Proof I have (real quotes, results, or "none yet"): [LIST]

Give me:
1) 3 positioning options. For each: target buyer, the situation where they'd choose me over each alternative, the competitor complaint it answers (quote it), and the main risk.
2) For each option, a one-sentence homepage headline and a one-sentence "unlike X, we..." line that only states things my bullets support.
3) A "do not claim" list: comparisons I might want to make that my evidence doesn't support.
4) One option you'd test first, and a cheap test (5 buyer conversations, a landing-page variant, or a cold email) to check it.

Rules: no superlatives ("best," "#1," "fastest") unless I can prove them. Keep competitor names out of any public copy suggestions unless the claim is a sourced fact.

Take the winning headline into the landing page copy guide, and check whether it changes what you can charge with the pricing guide.

Common traps

• Asking ChatGPT to research from scratch. You get a plausible list, not evidence. Paste real pages.
• Ignoring "do nothing." For many small tools, the biggest competitor is the buyer's current spreadsheet.
• Copying a competitor's feature list. Matching features rarely gives buyers a reason to switch. Answering a repeated complaint can.
• Undated prices. Write the date next to every price you record, and recheck before you publish a comparison.
• Trash-talking rivals in public copy. Unsourced claims about named competitors can backfire. Describe what you do instead.
• Doing it once. Put a quarterly reminder in your calendar and rerun Prompt 2 with fresh pages.

The short version

1. Collect real material: headlines, pricing pages with dates, and review quotes with links.
2. Use Prompt 1 to map every alternative, including workarounds and doing nothing.
3. Use Prompt 2 to build a grid from pasted sources only, with FACT, INFERENCE, and ASSUMPTION labels.
4. Use Prompt 3 to pick a positioning angle, a headline, and a cheap test.
5. Recheck and rerun every quarter.

ChatGPT can turn an afternoon of copy-pasting into a clear grid. The judgment about where you win still comes from what buyers tell you.