When employees skip the chatbot, asking a person still feels easier or safer. Look at where it lives, what it covers, what happens when it’s wrong, and whether people trust what they type into it.
Isn’t low use just a training problem?
Sometimes. Check the chatbot before blaming the people. The reasons below are things you can test.
1. Can people find it?
If it lives somewhere employees rarely go, they won’t think of it. Put it where they already work, and make it the same place every time.
Practical check: Ask someone outside HR to find it without help.
2. Does it cover what they actually ask?
A chatbot built around policies may not know about pay dates, forms or who to call. Compare what employees ask HR with what the chatbot can answer.
Practical check: Ask five employees what they’d ask it, before guessing.
3. What happens after one wrong answer?
One wrong answer can send people back to email for good. Showing the source for each answer, and making it easy to report a wrong one, helps people judge it for themselves.
4. Is there an easy way to reach a person?
If the route to HR is hidden, people go around the chatbot entirely. Make the handoff visible from the start.
5. Do people know who can see what they type?
Employees may hold back if they don’t know whether their questions are stored or read. Say plainly what happens to a conversation.
A fictional example.
- What they saw
- Few employees used the chatbot after launch.
- What they found
- It covered policies well, but most questions to HR were about pay dates and forms.
- What changed
- The pay calendar and the forms page were added as sources, and the chatbot moved to the page employees open every day.
Read more: Why an HR Chatbot Gives Bad Answers, and What to Check · How to Tell Whether an HR Chatbot Is Working
The example is fictional. This is general guidance, not advice for your situation.