Workforce advisor
TINIUN looks at the last 14 days and the next 7, 14 or 30, finds where you will be short and who is stretched, and turns it into a ranked list of what to do. The numbers are calculated, the chat is guarded, and the advisor advises: it never changes a schedule.
It reads schedules, daily results, quality scores and holidays for the teams a leader leads or manages. Open it from Advisor in the menu; it is on by default for the roles that can use it.
| Section | What it tells you |
|---|---|
| Coverage gaps | Hours inside a team's operating day with fewer people than its minimum, days a team usually works but nobody is scheduled, and who could take each gap without breaking a rule |
| Workload and burnout | Hours above 40 a week, days in a row, short rests, night, weekend, holiday and very long shifts, scored out of 100 with the reasons |
| Trends and projections | Adherence, attendance and shrinkage by day, with a projection and the date adherence is on course to cross its target |
| Quality | Each team's average score against the previous 30 days, failing evaluations and open disputes |
| Holiday cover | How many people each team has on each upcoming holiday against its minimum |
The list at the top turns findings into actions, Act now first: a gap and who could cover it, a person at high burnout risk and what drives it, people scheduled over the weekly limit, a quality drop, a holiday short of cover. Each item opens the page where you fix it, and the list exports to CSV for your huddle or ops review. Pair it with the scheduling board to act, and with Huddle to talk it through.
The score describes a schedule, not a person. Under 30 is low, 30 to 54 is the watch list and 55 and over is high. Each row shows reasons in words, such as "11 days in a row" or "3 short rests between shifts (shortest 6.5 h, minimum 8 h)". The limits for a short rest, a long shift or too many days in a row are the ones you set in your scheduling rules. Use it to rearrange hours or offer a day off, not to rate performance.
An optional chat lets leaders ask in plain language, for example "Where are we short next week?". It appears only when an organization administrator turns on the AI assistant.
| Automated | Stays with you |
|---|---|
| Calculating gaps, scores and projections | Deciding what to change |
| Ranking findings and suggesting who could cover | Editing and publishing the schedule |
| Explaining the figures in chat | Talking to the person, offering time off |
| Counting holiday cover | Approving requests and overtime |
FAQ
No. It advises. It lists findings and a ranked What to do list with links to the page where you make the change. It never edits a schedule, approves a request or sends a message, and the AI chat has no tools at all.
Every figure is calculated in TINIUN from schedules, daily results, quality scores and holidays, and is unit-tested. The AI explains the figures; it never produces them.
No. It describes a schedule: hours above 40 a week, days in a row, short rests between shifts, night, weekend and holiday shifts and very long shifts, combined into a score out of 100 with the reasons in plain words. Use it to start a conversation, rearrange hours or offer a day off, never as a performance rating.
A projection is a straight line through the last 45 days of daily results with a band of one standard error. It needs at least a week of history and says so when it does not have it. It cannot see holidays, promotions, outages or staffing changes, so treat it as an early warning, not a forecast.
Every message is screened before any AI is called for prompt injection, role or jailbreak attempts, hidden instructions, code and queries, requests for secrets or platform internals, other-organization requests and encoded payloads. A small model then checks the question is about workforce analysis at all and refuses if unsure. Replies are screened for anything that looks like a secret and cannot carry images or links outside TINIUN.
A leader sees the advisor for the teams they lead or manage. The chat answers from a snapshot of the same figures for those teams, never another organization.
Your question and the snapshot of figures needed to answer are processed by Anthropic to generate the reply. TINIUN does not store the conversation. See the Privacy Policy for details.
Through the MCP server, an assistant with the advisor scope can read an advisor summary that returns team-level counts and gaps without names, and use ready-made coverage-check and burnout-review prompts. It acts with the connecting person's own access.
We will load a realistic week and show the gaps, the watch list and the ranked actions.
Zendesk is a trademark of Zendesk, Inc.; TINIUN is not affiliated with or endorsed by Zendesk, Slack, or Microsoft. Slack and Microsoft Teams are trademarks of their respective owners.