HR Agent
Answers employee questions, screens applicants, and keeps leave, contracts and payroll straight.
HR teams spend an enormous share of their week answering the same forty questions and moving candidates between inboxes. The HR Agent handles both, and because it sits on the same database as payroll and time tracking, its answers are the actual answers rather than a link to a policy PDF.
Included on every plan. Switched on and scoped per department by an administrator in your own dashboard.
What it does.
Concrete work, not capabilities. This is what the agent is doing while your team gets on with something else.
-
Answers the questions employees are slightly embarrassed to ask
How much leave do I have left, when does my contract renew, what happens to my pension if I go part time, how do I claim these train tickets. It answers from that employee's own record, and only their own record.
-
Screens applicants against the role you actually wrote
It reads incoming CVs, compares them to the job opening in your system, and produces a ranked shortlist with the reasoning written out so you can disagree with it.
-
Catches the administration everybody forgets
Probation periods ending, contracts expiring, certifications lapsing, onboarding tasks not completed in week one, leave balances that will expire at year end.
-
Prepares payroll input and flags what looks wrong
Unapproved timesheets, overtime that breaks a threshold, a salary change that was never signed off, an expense claim sitting for three weeks.
- How many days of leave does Marieke have left this year?
- Which contracts expire in the next ninety days?
- Shortlist the applicants for the warehouse supervisor role and tell me why.
- What needs approving before I can run payroll?
What it will not do.
Limits are the useful part. An agent that could do anything would not be allowed anywhere near a ledger.
- An employee only ever sees their own record
- Salary and medical information stay behind the same permissions as the apps
- Never makes a hiring, disciplinary or termination decision
The applications it reads.
It can see these, through your permissions, and nothing else.
The rules that apply to this agent, and to all of them.
They remember nothing
An agent holds the conversation while you are having it and then lets go. Nothing you ask is stored as agent memory, added to a profile, or carried into tomorrow. The next question starts from your live data again, not from a summary of last week.
They read your business, not the internet
Every answer is assembled from the records in your own Site at the moment you ask. No scraped context, no vector store of your documents sitting somewhere else, no stale index quietly going out of date.
Your data never trains anything
Your records are not used to train, fine-tune or improve any model, ours or anyone else's. Trivena Technologies does not sell, share or aggregate customer data, and there is no product that would benefit from it.
They inherit your permissions exactly
An agent cannot see a single record the person asking cannot see. Ask about a salary you have no access to and the agent does not decline politely. It genuinely cannot read it.
