No. SleekAI does not connect to your ILS (Koha, Sierra, Polaris, Symphony) for patron account data. The bot cannot see holds, fines, or checkout history, and the instruction tells it to direct any account question to the patron's account login or the circulation desk. This keeps patron privacy intact and avoids any cross-system risk. The bot is a content guide, not a circulation terminal.

It can point patrons to relevant subject databases and digital resources, but it is not configured to act as a reference librarian. The instruction explicitly directs research-heavy questions (citation formatting, primary source location, in-depth subject research) to the reference desk or to the digital resources page. The bot is good at "where do I find X resource" and explicitly bad at pretending to be a research assistant.

If your library publishes a holiday schedule or maintains closure dates in WordPress, SleekAI reads them and answers accordingly. "Are you open Memorial Day" gets the published closure plus the next open date. If you don't publish closures (some smaller systems don't), the bot says so honestly and directs the patron to call the branch rather than guess. Honesty is what keeps the bot trustworthy across thousands of conversations.

Yes. SleekAI Multibot supports per-branch bots scoped by URL pattern or page parent. A central library bot might handle system-wide questions; branch bots can be scoped to that branch's hours, programs, and contacts. For a multi-branch system this keeps the chat focused without splitting your site into separate WordPress installs.

Yes. Because SleekAI uses general-purpose LLMs via your own API key (OpenAI, Anthropic, Google, or OpenRouter), the bot answers in whatever language the patron writes in. For library systems serving Spanish, Mandarin, Vietnamese, or Russian speaking communities, this is a significant access improvement at zero per-language infrastructure cost. Forms are still returned in whatever languages your site publishes them.

The bot is configured not to ask for patron names, addresses, library card numbers, or any account info. The guideline filter catches off-topic or out-of-scope prompts and steers them back. For library systems with strong child privacy policies (which is most of them), this means the bot can run on the kids' programming page without collecting any data that wouldn't be safe to collect on a public form.

If your room booking or pass reservation system has a WordPress front end, the bot can deep-link patrons to it with the right room or pass type pre-selected. The bot does not directly write reservations; that flow stays in your existing booking plugin or third-party system, with the bot acting as the conversational front door for the right page.

On your WordPress install, in the database, with model name, token usage, and page URL stored with each transcript. For libraries, this is also a useful programming intelligence tool: aggregate the questions and you discover which programs people are looking for, which days they're asking about, and which digital resources are underdiscovered. That feedback loop tends to improve event copy and digital resource page design over time.