01 Measure
Onboard a domain
Drop in your URL. Mentionwell crawls the homepage, sitemap, robots.txt, and structured data, then synthesizes a brand profile, audience, pitch rules, competitor set, content taxonomy, and a batch of starter headlines tuned to the questions AI engines are already answering.
02 Measure
Baseline the AEO scan
A buyer-intent prompt corpus runs across up to ten answer engines and the locales you care about. For every prompt Mentionwell captures the engine answer, the fan-out sub-queries, and every cited URL with position and source type — owned, competitor, or third-party.
03 Explain
Explain what won
Every cited page is scraped and fingerprinted: schema, headings, word count, byline, statistics, and claim-level facts. You see exactly what each engine lifted and why — not a single vanity score, but URL-level evidence and a share-of-voice trend you can stand behind in front of a CMO.
04 Act
Draft the gap-closer
The recommendations engine ranks the top actions per scan. For each, Mentionwell runs the multi-stage pipeline — research, outline, draft, editorial critic, metadata, FAQ, embedding, and image generation. Choose a normal draft or a Run + AEO draft that injects the scan evidence into the outline, draft, and FAQs.
05 Act
Review what needs a human
The editorial critic auto-rewrites anything below its confidence bar before the draft leaves the pipeline. With autopilot on, articles publish automatically — but any draft built on thin research or below the auto-publish confidence bar is held for your review rather than shipped unattended. Kanban gates let you add approval at any stage.
06 Act
Deliver to your stack
Articles serve from the public Reader API, fire as an HMAC-signed publish webhook, or commit as MDX to your repo. Each one ships with BlogPosting + FAQPage + BreadcrumbList JSON-LD, RSS, JSON Feed, sitemap entries, a per-article .md mirror, and a site-wide llms.txt — then pings IndexNow, Bing, and the Google Indexing API.