Answer employee policy questions from your own documents
An assistant answers questions about leave, expenses, benefits and conduct from your current policy documents, quotes the passage it relied on and sends anything unclear or personal to HR. For HR teams answering the same questions in chat all day.
Function: HR · Industry: Any industry
Answer employee policy questions from your own documents
6 steps
At a glance
- Trigger
- An employee asks a question in Slack, Teams or the intranet.
- Typical inputs
- Policy handbook and benefits documents
- Country and entity mapping for each employee
- Policy effective dates and versions
- Escalation rules for sensitive topics
- Systems it usually connects to
- Slack
- Microsoft Teams
- Confluence
- Google Drive
- How the work splits
3 rules, 2 model, 1 person steps
How it runs, step by step
6 steps, each tagged by what does the work. Rules are fixed logic, a model reads or drafts, and a person decides.
- Rules Fixed logic in code
- Model AI reads or drafts, with a confidence score
- Person A named person decides
- 1
Check who is asking
RulesThe employee's country, entity and level are looked up so that the right policy version applies.
- 2
Find the relevant passage
ModelA model searches the approved documents and selects the passages that answer the question.
- 3
Write a cited answer
ModelThe answer is written in plain language with a link and a quote from the policy, and says plainly when the documents do not cover it.
- 4
Apply the guardrails
RulesTopics on your sensitive list, such as grievances, medical leave details or pay disputes, are never answered and go straight to a person.
- 5
Answer the escalations
PersonAn HR partner replies to flagged questions and to anything the assistant could not answer.
- 6
Review gaps
RulesUnanswered and corrected questions are listed monthly so HR can fix unclear policies.
Where people approve
The decisions that stay with your team, whatever the confidence score says.
- HR owns and approves the documents the assistant may use.
- Sensitive topics are answered by people.
- Policy changes are published to the assistant by HR, with an effective date.
KPIs to track
What to measure, before and after. Targets come from your own baseline.
- 01Share of questions answered without HR
- 02Answers corrected by HR after the fact
- 03HR hours spent on repeat questions
- 04Questions that showed a gap in the policy
Questions about this workflow
What if two policies conflict?
The assistant shows both passages and tells the employee to confirm with HR. A conflict is also logged for the monthly review, because the fix belongs in the document.
Is employee data sent to a public model?
It depends on the deployment you choose. For HR data we usually recommend a private or enterprise-agreement model setup. We confirm your data rules before the build.
Which steps use rules, a model and a person?
Rules: check who is asking; apply the guardrails; review gaps. Model: find the relevant passage; write a cited answer. Person: answer the escalations. Rules are fixed logic in code. A model reads or drafts and reports its confidence. A named person decides.
What do we need to start?
Typically: policy handbook and benefits documents; country and entity mapping for each employee; policy effective dates and versions; escalation rules for sensitive topics. You also need read access to Slack, Microsoft Teams, Confluence and Google Drive, or the equivalents you use. A workflow audit measures your volumes and past cases first, then gives a fixed price for the build, so you can decide with real numbers.