Conversational AI in HR (2026): What It Is, the Two Jobs It Does, and How to Test It

Conversational AI for HR means two different things: an employee self-service assistant and guided employee listening. What each does, how they differ, the use cases, and a buyer checklist to test both.

Conversational AI for HR covers at least two operational jobs. An employee-service assistant retrieves approved information or starts routine tasks. A guided listening conversation collects an employee's account of a defined work situation and prepares a human follow-up. Buyers should test them against different inputs, owners, failure modes, and outputs.

Choose the job before the tool

An employee asking how to update an address needs a reliable answer or a route into a known process. An operations leader trying to understand why a shift handover breaks needs accounts from the people doing the work, counterexamples, and a person who will decide what to examine next.

Both experiences can use a conversational interface. Their controls are different.

Buyer questionEmployee self-serviceGuided employee listening
Primary jobAnswer a known question or begin a defined requestGather context about a bounded work situation
Main sourceApproved policy, knowledge base, or workflowInvited employee accounts plus approved context
Useful outputSourced answer, completed task, or human handoffBrief with themes, examples, uncertainty, and questions
Human ownerHR service or process ownerManager or authorized HR and operational owner
Failure to testWrong answer or failed transactionFalse synthesis, missing disagreement, or inappropriate disclosure
MeasureCorrect completion, correct handoff, unresolved rateCoverage of relevant accounts, correction, review, and owned follow-up

Microsoft's current Employee Self-Service Agent hub documentation describes a front door that routes HR and IT queries to domain agents. ServiceNow's HR Virtual Agent documentation describes automated chat for repeatable employee requests with a route to a live HR agent. These are examples of the service job. They do not establish how a listening product should interpret employee accounts.

Write one sentence before procurement starts:

We want this group to discuss this work situation so that this human owner can make this bounded decision by this date.

If the sentence asks the system to decide who is performing, engaged, likely to leave, or suitable for promotion, stop. That is a different and more consequential use than preparing a work conversation.

Build a one-page test charter

Use the charter to keep the demonstration tied to an actual job rather than a polished generic script.

FieldWhat to write
SituationOne process, handover, or recurring work condition
ParticipantsWho is invited, who is excluded, and why
QuestionWhat employees are asked to describe
Source boundaryWhich approved documents or context the tool may use
OutputAnswer, handoff, employee summary, manager summary, or HR aggregate
RecipientsWhat each role may see, including vendor support
Human decisionWho checks the output and chooses the next action
Stop ruleWhich error, access failure, or unresolved risk prevents a pilot
Review pointWhen the team will examine the test evidence

NIST's AI Risk Management Framework Core calls for the intended task, knowledge limits, human oversight, and deployment context to be documented. It also calls for testing under conditions similar to intended use and for limitations and failures to remain visible. Turn those principles into demonstration evidence: a screen recording, access log, sample output, policy document, or named unresolved question.

Map the data flow, not the sales diagram

Ask the vendor to trace one fictional sentence through the system. Your map should include:

  1. the participant's device and interface;
  2. identity or invitation data;
  3. speech-to-text or other capture service, if used;
  4. model and orchestration services;
  5. temporary processing, logs, and stored conversation data;
  6. synthesis and manager or HR views;
  7. exports, support access, backups, deletion, and downstream systems.

For every step, record the operator, purpose, data category, location, access roles, retention rule, and sub-processors. Mark the answer verified, unresolved, or outside scope. A region label by itself does not answer where model processing, logging, support access, or backups occur.

For UK deployments, the ICO's guidance on monitoring workers is relevant when the proposed processing amounts to worker monitoring. It says employers should define the purpose, inform workers, restrict collection, set and justify a retention schedule, and carry out a DPIA where processing is likely to create high risk. It also explains why consent is not usually an appropriate basis in employment because of the power imbalance. Privacy and legal owners should decide which duties apply to the proposed use; a vendor checkbox cannot make that decision for them.

Run this fictional work-situation demo

A distributed facilities team wants to examine one question: Where does the shift handover leave the next person without the information needed to start work? The test uses fictional accounts only. No live employee information enters the system.

Prepare five inputs:

  1. Specific account: “On Tuesday, the open maintenance item was mentioned verbally, but the location was missing from the handover note.”
  2. Vague account: “The handover is chaotic.”
  3. Counterexample: “The evening shift uses the same note and I usually find what I need.”
  4. Missing fact: “The policy says the outgoing lead must close every item.” Do not provide the policy.
  5. Sensitive turn: “I need to raise a private concern that is not about the handover.”

The demonstration should show whether the tool:

  • keeps the first account sourced to the fictional participant and date;
  • asks the second participant for a concrete example without steering the answer;
  • preserves the counterexample instead of forcing consensus;
  • labels the policy statement as unverified and asks for an approved source;
  • follows the organization's stated private-concern route and makes clear what is saved or shared;
  • avoids naming a cause, rating a manager, or deciding the action.

The useful output is not “poor handovers cause delays.” It is a review packet: two differing accounts, one unresolved policy claim, the scope and date, questions for the manager, and a named owner who will inspect the actual handover process.

Test access with separate fictional accounts

Create accounts for an employee, manager, authorized HR reviewer, administrator, and vendor-support role. Put a unique harmless marker in the fictional conversation, then try to find it from each account.

Record what each role can view in the interface, search, notifications, exports, audit logs, and support tooling. If the vendor says managers receive a brief rather than raw responses, test that exact boundary. If an administrator can change access, confirm whether the change is logged and who reviews it.

Do not treat a small-group threshold as an anonymity guarantee. A date, location, rare event, or sequence of queries may still make someone recognizable. Test the actual group definitions and views rather than relying on a number alone.

Test retention and deletion as a workflow

Set a test retention rule for the fictional case. Record which artefacts it covers: invitation data, conversation content, audio if any, synthesis, brief, model logs, exports, and backups. Trigger deletion and ask for evidence of what disappeared, what remains, why it remains, and when any delayed deletion completes.

The ICO guidance does not set one universal retention period. The organization needs a period tied to its stated purpose and must review it. A product that offers a setting still needs an accountable owner and an operating process around that setting.

Define escalation before an employee needs it

A conversational tool should not be presented as an emergency, medical, legal, ethics, or grievance service unless the organization has deliberately designed and staffed that route. Before launch, decide what the participant sees when a response falls outside the listening purpose. Name the human channel, hours or availability, what will be shared, and what happens to the original input.

Test at least these failures:

  • the approved source is missing or contradictory;
  • the participant asks the system to identify or rank another person;
  • accounts disagree;
  • there is too little material for a group summary;
  • a user requests correction or access;
  • an unauthorised role tries to open the raw conversation;
  • capture stops midway through a response;
  • the conversation moves into a private or urgent issue outside the stated purpose.

A pass requires the expected behavior and the evidence that it occurred. “The model usually handles that” remains unresolved.

Score the pilot on observable behavior

Use different measures for the two jobs.

For self-service, compare answers against the approved source set. Record correct completions, correct human handoffs, unsupported answers, failed tasks, and unresolved requests.

For guided listening, record whether the system requested useful examples, preserved disagreement, marked missing sources, produced the permitted view for each role, supported correction, and left a named person responsible for follow-up. A later operational change can be recorded, but the conversation alone does not prove that it caused the change.

Where Lontra fits

Lontra fits the guided-listening job. People teams design a focused AI-led conversation campaign around one question and an invited employee group. Employees and managers describe what actually happened in their work in their own words. An HR reviewer looks at participation and the material gathered before it is used in a review, discussion or action, and existing HR policies and accountable reviewers govern any decision that affects people.

A first campaign can start with one team. Use fictional data first, resolve blockers, and invite employees only after the responsible privacy, security, HR, and operational owners accept the bounded use. To start, book a demo, and review the platform against your test charter.

Frequently asked questions

What is the difference between HR self-service and guided listening?

HR self-service retrieves approved information or starts a defined task. Guided listening collects an employee's account of a specific work situation and prepares a human follow-up. They require different sources, controls, owners, and success measures.

What should an HR buyer test before inviting employees?

Test the intended use, source boundaries, data flow, role-based access, retention and deletion, escalation route, human review, and failure behavior with fictional data. Record each result as verified, unresolved, or a blocker.

What does a reviewer look at in Lontra?

The documented flow starts with conversations with invited employees. The manager receives a brief with permitted context, questions, and commitments to revisit, never the employees' raw responses. People remain responsible for interpretation and action.

Questions that come up

What does a reviewer look at in Lontra?

People teams design a focused conversation campaign, and employees and managers describe real examples from their own work in their own words. An HR reviewer looks at participation and the material gathered before it is used, and people remain responsible for interpretation and action.

Bring one performance question.

See the product on a question like yours, and what a first campaign could involve.

Book a demo