Twenty questions, approved by you
Claude drafts twenty buyer-intent questions across four types — best-of, comparison, problem-led, category-explainer. You review and edit every one before a single query runs.
Gun 02 · AI search analysis
SolvedAgain
Our own product, live and free to start. It answers one question with receipts: when an assistant recommends a brand in your category, is it you — and if not, why not?
The method
Four Steps, Nothing Invented
Every number on a SolvedAgain report traces back to a stored answer you can open and read. Volatility is shown, not hidden — mentioned in two of three Claude samples reads as two of three. Scroll on.
Asked with the web on
Every question goes to every engine on your tier, three times each, with live web search enabled. That measures the assistant your buyer is actually talking to — not stale training data.
Six metrics, one scorecard
Share of AI Voice, mention rate, average rank, recommendation rate, citation share, and volatility — against up to four competitors, broken out by engine, labelled grounded or not.
The reason you lost
For every question a competitor wins, SolvedAgain fetches the page the assistant actually cited and classifies the gap: content, format, authority, freshness, or schema. One gap type per loss, evidence attached.
Why it sits in this arsenal
Measurement Is Only Half a Volley
Most AI-search tools stop at a score. A score is not a plan.
- SolvedAgain names the gap — content, format, authority, freshness, schema — for every question you lose
- The owned front builds the pages that close a content or format gap
- The outreach front pursues the roundups, review sites and placements that close an authority gap — on the same engine that runs every other campaign
- The audit re-runs monthly with the same questions, so the delta report is proof the retainer moved something
SolvedAgain is a Broadside product. It runs at solvedagain.com and is free to start — no card required.