Draft, pending test: written from current vendor docs. We have not run this workflow end to end yet, so nothing below is a measured result. Not indexed.

AI competitor analysis: a step-by-step workflow with Grok and ChatGPT

Split the job: use ChatGPT (with browsing or Work) to read each competitor's own pages (homepage, pricing, changelog, docs) into a fixed comparison matrix with a source URL per cell, then use Grok to collect what real users say about them on X in the last 30 days. Merge both into one sheet, verify a sample of cells, and only then write the 'so what' for your positioning.

Who it's for & level

Founders, PMs and product marketers who need a competitor overview for positioning, a pitch deck or a launch, and want it sourced rather than copied from a chatbot's memory.

Level
intermediate
Tools
GrokChatGPTGoogle Sheets
Roles
Marketers Founders & small biz
Goals
Automate busywork
Tasks
Competitor research
Time to set up
60–120 min for a verified first version (est.)
Cost
Included in your ChatGPT and Grok/X plans; Google Sheets free. Longer agentic runs use more of your plan allowance.
Time saved
Half a day of manual tab-hopping for 5 competitors (est.) (est.)
Test status
Pending test
Last verified
pending test
Model versions
n/a
Written by
Usecased editorial (pending test) (agent)
Published
2026-10-15

Prerequisites

The problem

Asking a chatbot 'analyze my competitors' returns outdated facts from training data, mixed with invented features and no sources. Doing it by hand takes a day of tabs. You need a structured, sourced matrix plus a read of what customers say, fast enough to redo every quarter.

Steps

  1. Fix the matrix first. Decide columns before any AI run: positioning line, target customer, pricing (entry plan, price, billing), key features (5 you care about), integrations, recent changes, source URLs. Same columns for every competitor.
  2. Pass 1: their own pages (ChatGPT). Run the matrix prompt per competitor or for all, requiring a source URL per cell and 'not found' when a page doesn't say it. Prefer pricing, changelog and docs over the homepage.
  3. Pass 2: what users say (Grok). Ask Grok for X posts from the last 30 days about each competitor, grouped into praise and complaints, with links. This is the part their own pages won't tell you.
  4. Merge into one sheet. Paste both into the sheet: one tab for the matrix, one for user themes with links.
  5. Verify a sample. Open at least one source per competitor per column and check it. Anything that fails gets fixed by hand and the row flagged.
  6. Write the 'so what'. Now ask for gaps and positioning options, using only the verified sheet as input. You make the call.

Exact prompt & config

1. Matrix prompt (ChatGPT)
For each competitor below, visit their own site (homepage, /pricing, changelog or release notes, docs) and fill this table.
Columns: competitor | one-line positioning (their words) | target customer | entry plan name | entry price + billing | 5 features: [SSO, API, integrations, free tier, mobile app] yes/no | last visible product change (date) | source URL for each cell.
Rules: only facts from pages you opened in this run. If not found, write "not found". Don't use prior knowledge.
Competitors: acme.com, betaapp.io, gammahq.com
2. User-voice prompt (Grok)
Search X posts from the last 30 days about Acme (acme.com, @acme), BetaApp (@betaapp) and Gamma (@gammahq).
Exclude the companies' own accounts and obvious bots.
Per competitor: top 3 complaints and top 3 praises, each with 1–2 post links (author, date).
If fewer than 5 relevant posts, say so.
3. So-what prompt
Using only the verified sheet below (matrix + user themes), list:
1. Gaps no competitor covers well, with evidence rows.
2. Three positioning options for us, each with the competitor it's strongest against.
3. What we still don't know and how to find out.
Don't add facts that aren't in the sheet.

Sample output

Example output, written to show the expected shape. Not captured from a real run yet.

(Example of the expected shape, not a captured run)

| competitor | entry plan | price | SSO | last change | sources |
| Acme       | Starter    | $29/mo, annual | no (not on Starter) | 2026-09-30 changelog | acme.com/pricing, acme.com/changelog |
| BetaApp    | Free       | $0     | not found | not found | betaapp.io/pricing |
User themes (Grok): Acme — complaints: sync bugs (2 links), price increase (3 links)…

Why two passes and two tools

Competitors’ own pages tell you what they claim; users tell you what’s true in practice. One assistant rarely does both well. Splitting the work also makes verification easier: every cell has one source type.

The loop that keeps you in control

  1. Columns first, same for every competitor.
  2. Sources required per cell, ‘not found’ allowed.
  3. Verify a sample before analysis.
  4. Conclude only from the verified sheet.

Going further

Automate the pricing column weekly with our ChatGPT Work price-monitoring case, and run the Grok pass monthly. Feed recurring complaints into your positioning and our social media angles workflow.

How it went

Pending test. What we expect, based on the vendor docs and public reports below. This becomes a measured result after our run.

Expected result, pending our test: a usable matrix in well under an hour with some 'not found' cells and a few wrong cells caught in verification, plus user themes that add things competitors' pages don't show. We'll record cell-level accuracy on the sample we check and total time.

Worth knowing

FAQ

Can AI do a competitor analysis?

It can collect and structure public information fast, if you require sources per fact and verify a sample. The conclusions should stay with you.

Which AI is best for competitor analysis?

Use a browsing assistant (ChatGPT) for competitors' own pages and Grok for real-time X reactions. Each covers what the other can't see well.

How often should I update a competitor analysis?

Quarterly for the full matrix; weekly automated checks for pricing pages if price changes matter to you.

Related use cases