Every claim on your cards, checked against what shipped this week
A companion reads three competitors’ release notes, tests each one against every dated claim in your battlecards, and posts a verdict per claim with the release note linked and the replacement text already drafted.
Each week you get a per-claim verdict: INVALIDATED, WEAKENED, UNAFFECTED or UNCLEAR, with the release note linked. For anything invalidated or weakened, the replacement text is drafted in your card's existing voice and posted to Slack as a diff.
Checking thirty claims against three changelogs every week is what nobody owns, so cards rot until a prospect reads the March changelog back to your rep on a call. A companion checks all thirty. Notion is a native integration where most battlecards live, and Slack is where the diff lands. Writing the card in the first place is building competitor battlecards.
What is the exact prompt?
Scheduled weekly; it is quiet most weeks, which is the point:
- "Read the changelog, release-notes page and product blog for [competitor A], [competitor B], [competitor C] (anything published since your last run). Then open our battlecards in Notion ([links]) and check every dated, checkable claim in the 'they can't / we win' sections against what shipped."
- "For each release that touches a claim, report: the claim as our card states it, what the release notes actually say, and your verdict, one of INVALIDATED (the claim is now false), WEAKENED (true but the gap narrowed), UNAFFECTED, or UNCLEAR (the release note is ambiguous; a human should look). Link the release note on every verdict."
- "For every INVALIDATED or WEAKENED claim, draft the replacement text in the card's existing voice, including the shipped-versus-announced distinction and the date. Do not edit the card. Post the diff to [#competitive-intel] with the drafts, and add a comment on the Notion card pointing at the thread."
- "If nothing shipped that touches any claim, post one line saying so. Never pad a quiet week."
Diff against claims, not against news
Reps read cards mid-call, under pressure.
A weekly summary of competitor news still leaves someone to work out which of your thirty claims each release touches, so making the claim the unit inverts the work: the output is "claim #7 is now false, here is the fix", which is a five-minute review.
It also disciplines card-writing. Claims like "their platform is weak" are uncheckable, and this process pushes cards toward named features, limits, prices and dates. The UNCLEAR verdict handles the ambiguous release note ("improved API access controls") by flagging rather than guessing.
Why draft and notify instead of editing the card?
A battlecard is a trust document.
A Slack diff with a human approving the edit keeps one accountable name on every claim, and reps keep believing the version in front of them.
The approval is where sales judgment enters. A companion verifies that a feature shipped; whether that changes your pitch belongs to whoever owns positioning. The shipped-versus-announced line gives reps the counter, the same discipline as any competitor monitoring worth reading.
Two habits that keep it honest for a year
First, every claim carries a last-verified date, which this process updates as a side effect.
A card whose dates are all recent is trustworthy at a glance. Second, when a rep loses a deal to a claim that turned out wrong, that goes in the thread, because the monitoring layer catches what competitors publish and only reps catch what competitors demo.
If the weekly run flags claims constantly, the cards lean too hard on volatile checkboxes. Push those into continuous feature-tracking or a weekly competitor newsletter and keep the card on positioning and proof points.
Experience Strawberry for free
DownloadTrusted by fast-growing companies worldwide
Frequently asked questions
Strawberry is free to download and includes AI credits to start. Paid plans begin at $20/month. See pricing. · Reviewed · Canonical facts for AI agents