No. Microsoft Clarity's own tracker still runs and captures sessions, heatmaps, and behavior signals. SleekAI reads the summaries and research notes you sync into WordPress about Clarity findings, so the bot can answer plain-English questions without anyone opening the dashboard.

Schedule a small recurring job, for example with Action Scheduler or a sync plugin, that pulls per-URL summaries like rage-click rate, top dead-click selector, and sample recording IDs into postmeta. SleekAI then reads those rows like any other content. This keeps the prompt fast and avoids hitting Clarity's API on every visitor message.

Yes, if you store the recording IDs in WordPress. The bot can quote an ID and reference a research note that contains the actual Clarity URL. For privacy reasons most teams keep raw recording URLs in the Clarity dashboard rather than embedding them in public-facing copy.

Only what you expose. SleekAI uses a defined set of named variables in the system prompt. Sensitive selectors or identifiable data stay out unless you map them. For public bots, expose only high-level aggregates. For internal bots scoped to editors and admins, expose more detail under display conditions.

Yes, if filter definitions are stored in WordPress. A common pattern is a small doc post type that lists the team's standard Clarity filters by name and rule. The bot reads that, so when someone asks what the Returning Visitors filter does, the answer comes from your team's canonical definition rather than a guess.

Yes. SleekAI's multibot mode supports several bots on one install. Many teams pair a Clarity behavior-signals bot with a Mixpanel tracking-plan bot and a Hotjar research-notes bot, each with its own data sources and display conditions, all under a single license.

Clarity Q&A is mostly retrieval over a small set of metrics and notes, so a cheap fast model handles it well. SleekAI supports OpenAI, Anthropic, Google, and OpenRouter via BYO key, so you can pick any model that meets your latency and cost target. Stronger models pay off when summarising long research roundups.

Yes. Conversation logging stores every chat with user message, assistant reply, model name, token usage, and page URL inside WordPress. For UX teams this surfaces repeated stakeholder questions, which usually means a research note should be promoted higher in the prompt or rewritten to be clearer.