Skip to main content
Beebole AI is the set of AI features built into your Beebole account: time entries suggested from your own activity, reports built from a plain-language question, and an extra pair of eyes when you review a submitted timesheet. Everything lives on the Assistant page — click Assistant in the sidebar when AI features are enabled for your account.
AI features run on Beebole-operated servers. Nothing is sent to third-party AI providers.

Suggested time entries

Beebole drafts time entries for you and shows them in a Suggested entries tray above your timesheet. Suggestions come from several sources: Suggestions are private — only you see them until you accept. For each one you can:
  • Accept it (or Accept all) — the entry lands on your timesheet.
  • Dismiss it.
  • Click Why? to see the activity behind a desktop suggestion. If you’re on a different machine, Beebole tells you: These details aren’t available on this device.
In the timesheet’s calendar view, suggestions also appear as ghost entries: drag a suggestion to a time slot to schedule it, or accept it as-is.
Activity captured by the desktop app stays on that device, is never shared with anyone, and is deleted after a week. Only the entries you accept reach your timesheet.

Report builder

Describe the report you need in plain language and Beebole creates and runs it for you.
  • On the Assistant page, use Ask for a report — the report opens ready to run.
  • On the Reports page, type what you want to know in the report builder.
The result is a regular saved report: refine its filters, move it to a folder, or delete it like any other. See Reports. The report builder appears only for roles allowed to manage reports.

Approval review

When you review a submitted timesheet, anything unusual is flagged before you decide. The review digest checks the period for:
  • Time on a non-working day
  • More time than scheduled
  • Time outside working hours
  • Time on an archived project / Time on an archived task
  • Entry right after the lock date
  • Task over its planned time
  • Period total far from usual
Approvers also receive an email digest summarizing the timesheets awaiting their review. See Approval workflows.

Already working for you

Some AI touches need no setup at all: timesheet reminders mention how many suggested entries are waiting, and budget alerts include a projected overrun date based on the current burn rate (see the Budget Status report).

Connect your own AI tools

Beebole also works with the AI assistant you already use. Connect Claude, ChatGPT, or another assistant to your Beebole data — with exactly your permissions — and manage every connection from the Connected apps list. See the Beebole MCP server for setup.

MCP server

Connect Claude, ChatGPT, and other assistants to your Beebole data.

Desktop app

Draft time entries automatically from what you work on.

Browser extension

Turn time on the sites you choose into draft entries.

Timesheets

Where suggested entries appear and get accepted.

Frequently asked questions

No. Beebole’s AI features run on Beebole-operated servers, and nothing is sent to third-party AI providers. Connecting your own assistant through Beebole’s MCP server is the opposite direction — your assistant reads your Beebole data with your permissions, and only when you set that connection up yourself.
Only you. Suggested entries in Beebole are private to you until you accept them — teammates, managers, and admins never see your suggestions or the activity behind them.
From your own recurring logging patterns, from the Beebole desktop app and browser extension if you install them, and from Kanban cards you move to a done column. Each suggestion shows its source, and you decide what to accept or dismiss.
Describe what you want to know — on the Assistant page under Ask for a report, or directly on the Reports page. Beebole builds the report and runs it; the result is a normal saved report you can refine and reuse.
It flags time on non-working days, overtime beyond the schedule, time outside working hours, entries on archived projects or tasks, entries right after the lock date, tasks over their planned time, and period totals far from that person’s usual — all before you approve or reject.