Yes. SleekAI maps any CPT, taxonomy, and ACF field to bot variables. Whether your stories live under case_study, customer_story, or default Posts with a category filter, the bot reads the same content your /case-studies/ archive does. You configure the mapping once in the SleekAI admin and the bot reads live data on every request.

The system prompt forbids inventing customers. The bot will say something like 'we do not have a published story in that exact segment yet, but the closest match is X' and then offer to route the visitor to your sales team. This protects the credibility of every other answer it gives and stops the obvious sales-killing failure mode of fabricated references.

Yes. Numeric ACF fields become structured variables the model can reason about. A query like 'show me stories with payback under 6 months' is answered by comparing the payback_months field across the corpus, not by guessing from the headline. Range filters, minimums, and maximums all work cleanly.

Every reply ends with a link to the matching /case-studies/{slug}/ page. This is enforced in the system prompt and reinforced by the demo conversations. Buyers can verify the claim, scan the rest of the story, and download any PDF you offer, while you get clean analytics on which case studies drove conversions.

The bot can summarise the publicly visible portion (often the title, customer name, industry, and headline metric) and route the visitor to your existing gating form for the full PDF. You do not have to ungate the story, and the bot does not leak gated content. It also passes UTM parameters through to the form so attribution stays intact.

Site search expects the visitor to know the right noun. A buyer asking 'who else like us solved this' rarely uses your taxonomy names. The bot translates natural-language intent into the right facet query, ranks results by relevance plus metric strength, and replies with prose rather than a list of links. It works alongside search rather than replacing it.

Yes. Multibot lets you scope one bot to /case-studies/ for buyer research, another to /partners/ for channel-partner-facing stories, and a third to /investors/ for outcome-heavy summaries. Each bot can have its own system prompt, presets, and display conditions, and they share the underlying CPT and ACF data.

Every conversation is stored with origin page, model used, token usage, and a transcript. Marketing can mine the logs for unmet demand (stories visitors ask for that you have not yet published) and sales can hand the transcript to the AE on a meeting handoff. Retention and export controls let you align logs with your data policy.