How can I use AI to help me understand my competitors?
- Type
- Blog post
- Published
- Nov 16, 2026
- Author
- Adam Warburton Co-founder and CPTO
- Reading time
- 6 min
Ask ChatGPT who your competitors are and you’ll get an answer in seconds. It might even be a decent one. But every answer is a net cast at random into a body of water you can’t see: where it lands depends almost entirely on how you worded the prompt; you never find out what it missed. Word the question slightly differently tomorrow and you’ll haul up a completely different catch.
That, in short, is the answer to the question in the title. General AI tools such as ChatGPT, Claude and Copilot can tell you who your competitors are and what they appear to be doing today. What they can’t tell you is how that picture is changing, why it’s changing or what you should do about it. For that, you need AI built on a history of product data: AI that turns movement into recommended actions.
What can ChatGPT, Claude or Copilot tell me about my competitors?
Ask a general AI model about your competitors and you’ll get a reasonable snapshot: who they are, roughly what they sell and what they appear to be doing right now. Depending on what it can reach, it might pull back a current price, whether a product is on promotion or what its reviews say today.
That’s a useful starting point. But the quality of the snapshot depends on things you can’t fully see: how well you’ve written the prompt; how capable the model is and how hard it’s reasoning; which retailers and sites it can actually reach. You rarely know how wide the net was. A competitor that wasn’t on the pages it happened to read simply doesn’t exist, as far as the answer is concerned.
Can ChatGPT track competitor prices?
Not in the way you’d need it to. It can look up a price if it can reach the page; these days it can also remember past conversations and even run a check on a schedule. But looking up is not tracking. Each check is another cast of the net; it might land on a different retailer, a different listing or a different page of results, depending on how the prompt was worded that day; memory holds onto whatever it happened to catch; it doesn’t make the catches comparable.
So you end up with a pile of answers rather than a record. There’s nothing reliable on when a promotion started or ended, nor any clear way of spotting that a rival has quietly cut prices three times in six weeks.
You get the immediacy of whatever is in front of the model today. Nothing more.
Why isn’t a snapshot of my competitors enough?
I posted about this on LinkedIn a few weeks ago and it’s a story I keep coming back to.
For nineteen days at the end of the summer, the best-selling product in cookware was a £4.99 egg poacher. It had only two reviews, but both were five stars. Then the rating slipped to 2.8, then to 1.9; one review reported that it had exploded all over the microwave. It was still number one.
Within a week, it hadn’t only lost the top spot: it had dropped to position 45, with two stars from 266 ratings.
So if you’d asked AI to name your biggest competitor in that category on any one of those nineteen days, it would have pointed you at a product that was, quite literally, about to explode. Ask it the day after the fall and the egg poacher wouldn’t have registered as a competitor at all, despite being number one two days earlier.
The signal was there all along; it just wasn’t in the snapshot. Follow the trajectory of the reviews alongside the rank and you’d have seen a flurry of sales running straight into a wall of poor ratings. That’s the difference between knowing what’s in front of you and understanding where it’s heading.
How does AI decide who counts as a competitor?
This is the question that keeps me up at night. Take an eight-seater dining table with a wooden frame and an aluminium top. Is a six-seater all-wood table a competitor?
Ask a general AI model and the answer is usually no: different configuration, different material, different product. Ask a shopper walking into a showroom and the answer is obviously yes. They’re both dining tables; the shopper would happily choose between them.
Competition is defined by the decision the customer is making, not by a matching spec sheet. A tool that can’t see that will hand you a tidy competitor list that misses half the products you’re actually up against. That’s why Rove scores how genuine each competitor match is, rather than relying on a spec sheet to decide.
What does AI competitor tracking look like when it’s built for the job?
Rove’s answer is history. We build an incredibly rich record of product data: movements in rank, price, promotion, ratings and reviews, tracked over time rather than captured once. The tracker keeps running, daily or weekly, so the picture sharpens the longer it runs.
That history lets us do two things a one-off answer can’t. First, we can tell you what’s happening with your competitors and what’s moving. Second, because we’ve modelled what happened in all of those previous moments, we can tell you what to do in response.
What should I do when a competitor drops their price?
Say you’re a brand and I tell you your closest competitor has just dropped their price by 5%. The instinct is to match it straight away.
The reality is that no brand exists in a vacuum. Your nearest competitor might have dropped price because they’ve just had two negative reviews. The competitor two places down might have put their price up by 10%. The one three places away might be out of stock; the one four places away might have quietly moved delivery from next day to three days.
That isn’t a single, one-shot view. It’s a matrix of moves happening at once, and our actions layer models that permutation to tell you what to do about it. Sometimes the right answer is to match the price; sometimes it’s to hold your nerve.
What do AI-recommended competitor actions look like?
The scope isn’t limited to two or three things, but the most common actions fall into a handful of areas:
- Price: when to move, by how much or when to hold.
- Promotion: whether a promotion is worth running, given what’s happening around you.
- Description and positioning: how your product is described and positioned against the competitors shoppers actually compare you with.
- Product configuration and composition: responding to what reviews are saying, about your product and theirs, in the product itself.
- Keep an eye on: one of my favourites. Sometimes a competitor is building momentum and you don’t need to do anything right now; you just need fair warning to watch it closely.
Actions delivered, not data dumped. That’s the point.