Your catalog isn’t one AI-search problem. It’s one for every product line.

Yes — that is how most catalogs are set up. A Project is its own monitoring context, and one workspace holds as many as your catalog needs. The example on this page runs four: Complexion, Lips, Skincare, Eyes, 53 tracked buying questions between them.

Yes. Tracked questions belong to the Project you add them to and appear nowhere else, so a mascara question never turns up in your skincare reading. You can build each Project's library from scratch or start from a shared starter pack and edit from there.

Yes, and they usually should. Competitors are set per Project, not per account, because the brands competing for a foundation answer are rarely the brands competing for a mascara answer. Each Project's competitive reporting is computed against its own set.

Yes. For every tracked question Sourcepath records what each monitored engine answered, which organisations were named in that answer, and which pages it cited. What you get is the observed answer and its sources — not a guess about why an engine chose what it chose.

They still show up. Any site cited often enough in your Project's answers is surfaced as a discovered organisation alongside the ones you track. It is identified by its domain, because a domain is what was actually observed — Sourcepath will not invent a company name from a hostname. In the Eyes example, three of the most-cited names were sites nobody had added.

Yes. Pages are fetched and analyzed for real structural signals — headings and their depth, FAQ sections and FAQ schema, comparison tables, content depth, internal linking and structured data — and compared against the pages AI actually cited for the same question. The recommendations you get are specific changes attached to the part of the page they belong to.

Yes. Sourcepath groups the cited pages by destination and classifies them — buying guides, review platforms, media, community, marketplaces and more — then shows how often each recurs across that Project's questions, and checks the page itself for your brand and your competitors by name. Each one routes to the marketing workstream it belongs to.

Yes. Every Project keeps its own scan history, so you can read the same Project at two points and compare like for like. Comparisons run over the questions that completed in both scans, and every figure carries the readings behind it — 6 of 44, not a percentage on its own. See how Sourcepath reports on that.

Yes. Projects sit side by side under one organisation, each with its own questions, competitors, scans and findings, and the Project Library reads all of them at once. Worth knowing up front: access is granted at the organisation level, so everyone in the workspace can see every Project — there are no per-Project permissions today.

Separate the catalog into Projects — one per product line — so each carries its own shopper questions, its own competitors and its own evidence. Sourcepath then shows, per product line, which brands AI names, which pages it cites, what your product and category pages are missing, and which outside lists keep shaping the answer. See how SEO and AEO teams run this.