Ask Hera
Ask Hera is the conversational side of the HR Tool's AI. You type a request in plain language — the prompt reads "Ask a question, or tell it what to do…" — and Hera either answers it or files it for approval.
That split is the thing to hold onto:
| You ask it to | What happens |
|---|---|
| Read something — a leave balance, a policy, an employee detail | Answered straight away in the chat, from your real HR data, with the records it used |
| Change something — approve a leave request, add a candidate or an employee, chase an onboarding item, sync to Xero | Filed as a proposal in the approval Inbox, for an HR admin to approve |
Hera never changes your HR data on its own. A request to change something becomes a proposal like any agent's, with the same approval step and the same audit trail.
Next to the title, Hera shows its status — On, Off, or Temporarily unavailable. If it's off, HR admins see a Turn on Hera button to switch it on for the whole company; everyone else is told to ask their HR team or administrator.
Asking a question
Type into the box at the bottom — its prompt reads "Ask a question, or tell it what to do…" — and send.
If you're not sure where to start, Hera offers suggested questions such as:
- How much leave do I have left?
- Show my upcoming shifts
- Show my pending approvals
- What else needs my attention today?
Reading an answer
Hera doesn't just reply in prose. Depending on the question, the answer comes back with supporting panels:
| Panel | What it shows |
|---|---|
| Sources | The records the answer was drawn from — check this before acting on an answer |
| Key points | The answer reduced to its essentials |
| Your leave balances | Live balances, when the question is about leave |
| Status | The state of the thing you asked about |
| Type, Dates, Employee | Details of the specific record, where one applies |
Sources is the panel that makes an answer verifiable. If an answer will drive a decision — approving something, telling an employee a figure — open Sources and confirm it came from the record you expect.
Conversations
| Control | What it does |
|---|---|
| New chat | Starts a fresh conversation |
| Recent chats | Your earlier conversations. Before you've had any it reads "Start a conversation to see history here." |
Attaching a CV
The composer accepts a file via Attach a résumé (PDF, DOC, DOCX), with Remove attachment to clear it.
Whatever Hera reads from an attached résumé stays available for the rest of that conversation, so you can keep asking about it in later messages. Attaching a new résumé replaces what Hera remembers from the previous one.
Attaching a CV and asking Hera to turn it into a candidate or an employee record is still being worked on — it is not a finished feature yet. Expect uneven results until it is announced as ready, and keep creating candidates and employees the normal way in the meantime. See Candidates Management and Employee Management.
If an answer fails
Hera will occasionally return "Sorry, I was unable to get a response. Please try again.", "Sorry, something went wrong. Please try again.", or, if something failed unexpectedly while it was working on your request, "Sorry, I ran into a problem (reference {reference}). This has been logged, please try again in a moment." — the reference is a short code support can use to look up what happened, instead of showing the raw error detail.
All three mean the request did not complete — nothing was changed in your HR data. Send the question again.
If every AI provider behind Hera is unreachable, you'll instead see a plain notice — "⚠️ I can't reach my AI service right now, so I can't look anything up in your live HR data this turn. Please try again in a few minutes — if it keeps saying this, let your ShiftCare administrator know." — followed by pointers to where Leave Management, Performance Review, Org Chart, Contracts & Documents and Training Management live, so you're not stuck while the AI is down.
Reporting a problem or sharing feedback
Two links sit above the composer: Report a problem and Share feedback. Either opens a short form without leaving the conversation.
| Field | Shown for |
|---|---|
| Report type | Both — Support request or Feedback |
| Summary | Both — up to 120 characters |
| What happened? / What would you like to share? | Both — up to 3,000 characters |
| What did you expect to happen? | Support request only |
| When did it happen? (include your time zone) | Support request only |
Click Review report to see exactly what will be sent — your company, ShiftCare account, and name, alongside the report itself. Only these details and a reference number go to support; chat messages, attachments, and browser diagnostics are not included.
Click Send report to submit. A successful report shows a Reference and, once delivered, a support Ticket number. You can send up to 10 reports in a 24-hour period; past that, sending shows a message asking you to try again later.
If support reporting hasn't been turned on for your company yet, the form reads "Support reporting is not available for this account yet. Contact your HR administrator for help."
What powers it
A Powered by marker in the header shows the engine. Hovering it reads "This company has connected Hera to Xagent" — all Phase 5 AI, Hera included, runs on Xagent.
- AI Agents — the agents that propose changes, and the approval Inbox
- People — the AI-side org view
- What's New: Phase 5 — the release summary