Stephen Turner
Akash Kumar
Jordan Lee
Taylor Scott
Use an AI agent to track competitor mentions on calls
Customers tell your sales team which competitors they're looking at, and what they heard about them. An after-call agent can collect those mentions with exact quotes. A person then decides what's true, since a customer's remark is a lead, not a fact.
The job: collect what customers say
Someone on a call says "we're also looking at Acme, and they told us they can do it for less". That line matters to product marketing, but it can easily stay in the rep's notes.
This agent does one narrow job. After a customer call it can see, it lists each competitor mention, with the quote and its context. It doesn't research the competitor or decide whether the claim is true.
Inputs, output and what it must not do
| Part | What it should be |
|---|---|
| Input | One completed meeting's recap and transcript |
| Reference | Your list of competitor names and common misspellings, in the instructions |
| Output | A short list in Slack, or in the Assistant chat while you test |
| Must not | Add facts from the web, score a competitor, or edit sales collateral |
Instructions you can adapt
This is a starting point, not a built-in template. Replace the names with your real list, and test it on calls you remember well.
- If this wasn't a call with a customer or prospect, stop and post nothing.
- Look for these names: Acme, Globex, Initech. Include clear misspellings, since transcripts can mishear names.
- If none appear, stop and post nothing.
- For each mention, give the exact quote, who said it, and whether it was the customer or our team.
- Label it: evaluating, currently using, used before, or passing mention.
- Add the topic, such as price, a feature or support.
- Never add facts from outside this meeting. Treat what a speaker says about a competitor as their claim, not as fact.
- If you have no exact quote, leave it out.
- Post the list to the #competitor-mentions Slack channel.
Set it up in Bridge
Create the agent in four steps: Identity, Abilities, Tools and Triggers. Paste the instructions in Identity, up to 4,000 characters. Reading transcripts is always on.
In Tools, leave web search off, since this job uses only the meeting. Connect Slack first, then turn it on so the agent can post. In Triggers, choose After every meeting. Custom agents and agentic workflows need a paid plan.
Keep a person in charge
The trigger runs for every meeting, which is why the first instruction is a stop rule. Test it on a rep's call to confirm which meetings it covers, and name one person, perhaps in product marketing, to read the list each week.
That person opens each quote in the transcript, and checks any competitor claim against the competitor's own site before it reaches a battlecard. Repeating a customer's claim as fact can embarrass your team, or mislead a buyer.
Prefer to see results before anything is posted? Keep Slack off and After every meeting off, then @mention the agent in the Assistant on a completed call.
Example output
An invented example, from a renewal call:
- Competitor: Acme.
- Quote: "Acme told us they include onboarding for free."
- Said by: the customer.
- Label: evaluating.
- Topic: price and onboarding.
Limits to plan for
- It only sees captured calls. Conversations in email or chat won't appear.
- It reads one meeting at a time, so it doesn't count trends. Tally mentions in your own sheet, or use a scheduled agent as a separate experiment.
- A transcript can mishear a company name, so the reviewer rereads the transcript, and any recording, on anything important.
- One customer's opinion isn't a market signal. Look for repeated patterns before changing your messaging.