Best tool to compare my brand vs competitors in ChatGPT: what we use for clients
How we compare a client's brand against competitors in ChatGPT answers: the prompt set, the share of voice views that matter, and the tool we run the work on.
Every client wants the same screenshot: ChatGPT recommending them over the competitor. What they need is something less photogenic, a fair comparison across many prompts that shows where they lose and why. Here is how we build that comparison and what we look for in a tool.
Start with the prompt set, not the tool
The comparison is only as good as the questions behind it. We write the set with the client in a working session and sort it into three types: category questions ("best CRM for small agencies"), brand questions ("is Brand X good for teams"), and head-to-head comparisons ("Brand X vs Brand Y"). Each prompt gets a topic tag and a country. We keep it fixed for at least a month. Changing prompts mid-stream makes any trend meaningless.
We also add the competitors the client names, then check which others show up in answers anyway. There is usually one the client forgot, and it is often the one winning the most prompts.
The views that matter
Averages mislead here. A client can show a decent overall share of voice while a rival owns the whole pricing topic. These are the views we rely on:
- Share of voice by topic. It shows which subjects a rival dominates.
- Position in the answer. First mention and fifth mention are different outcomes.
- A heatmap of brands against topics or models. One picture often replaces a page of explanation. Promptwatch's heatmap compares any two dimensions, so we use it in nearly every review.
- Period comparison. Before and after a content change, with the same prompts.
We run this in Promptwatch for clients. ChatGPT is monitored in its live search experience, prompt panels run daily, and competitors are auto-discovered, which catches the forgotten rival. Brand merging keeps two spellings of a competitor from splitting the numbers, and the Brand Book keeps mentions clean.
Explaining why the competitor wins
A comparison that ends at "they beat you" is not a deliverable. The follow-up is source analysis. Citation analytics shows which pages and domains ChatGPT cites for each prompt. Page Tracker lets us follow a competitor's specific URL and see how often it is cited. Usually the pattern is clear: their pricing page is current and specific, their comparison page answers the question in the first lines, and third-party sites carry reviews of them but not of us.
The content gap analysis turns this into work, with coverage scores per topic. Findings land on the Action Board as prioritized tasks we can assign. When a client is on a Business plan, the Content Agent can draft against the gaps, and we keep a review step before anything is published.
Tracking whether it paid off
Visibility without visits is a vanity metric. Visitor analytics attributes AI-referred traffic and conversions, so we can show whether a gain in share of voice brought people to the site. Crawler logs are worth a look too: if the bots rarely fetch a key page, no amount of editing will move that prompt.
Where other tools fit
We have looked at the alternatives. AthenaHQ has a solid action workflow, though its credit-based billing needs planning. Scrunch AI brings audits and bot traffic reporting. Profound is an enterprise agent suite. Suite add-ons from Semrush and Ahrefs are handy for directional checks if you already pay for the base product. Otterly.AI and Peec AI are quick to set up for a first look. For client programs we need the comparison, the sources and the traffic in the same place, so that is where we landed.
Setting expectations
We tell clients that ChatGPT answers vary run to run and by user, so no tool can promise a recommendation. What the comparison gives you is a stable view of where you stand against named rivals and a ranked list of fixes. Review it weekly, change a few pages at a time, and let the trend, not any single answer, tell you whether it worked.