No. SleekAI reads the summaries Site Kit caches inside WordPress on its scheduled refresh. That covers Search Console queries and pages, GA4 sessions and channels, and PageSpeed scores at the depth Site Kit stores them. For anything deeper or live, Search Console itself and GA4 remain the right tools; the bot quotes Site Kit's last-cache timestamp so freshness is explicit.

Not unless you configure it that way. Display conditions on user role, capability, logged-in state, and URL pattern keep the analytics bot scoped to administrators or editors inside wp-admin. Any public chatbot you run on the same site is a separate bot under multibot, with its own data sources and no access to Site Kit cache.

Yes, partially. If only Search Console is connected through Site Kit, the bot can answer search-related questions and skips GA4 queries. The system message includes which modules are active per the Site Kit settings, so the model knows which lookups to attempt and which to decline.

By default SleekAI is read-only against Site Kit cache. It surfaces what is cached and points at the dashboard for a manual refresh. If you want a chat-initiated refresh you can wire a custom PHP callback, but most teams keep Google OAuth and module setup inside the Site Kit UI where the consent flow belongs.

Yes. Conversation logging is on by default per chatbot, with user messages, assistant replies, model name, token usage, and page URL stored inside WordPress. For an analytics bot, the audit trail doubles as a feedback signal: the editorial team can see which queries keep coming up and write content for them.

Only if Site Kit cached both snapshots locally, which covers the default week-over-week and month-over-month comparisons on the dashboard. For arbitrary date ranges that Site Kit does not cache, the bot has to say so honestly and point at Search Console or GA4 directly rather than invent figures.

A fast model handles routine reads of cached queries and pages well. For longer prompts that synthesise search performance with GA4 channel trends and PageSpeed regressions, a stronger model from OpenAI, Anthropic, Google, or OpenRouter gives better synthesis. SleekAI lets you pick the model per chatbot.

Each SleekAI install runs its own chatbots with its own data sources, so configuration is per site. You can export the chatbot config and re-point the data sources on a new site, which is useful when the same Site Kit helper bot is wanted across a portfolio of client installs.