Chatbot Luis

Our nudge is called “Answer-first instead of Link-first”. Rather than responding to HR questions mainly with long documents or links, Luis first guides employees to the relevant information. Each response follows a clear sequence: first, the employee receives a short and understandable answer to the question. Luis then states the concrete next step. Only afterwards does he provide additional details, policies or links. If important information is missing, Luis asks a targeted follow-up question instead of generating a very general response. If he cannot fully resolve the issue, he prepares the request so that it can be handed over to HR easily.

Accept the cookie setting to see Panopto content.

Unlock Panopto

What does the topic mean?

Our nudge is called “Answer-first instead of Link-first”. Rather than responding to HR questions mainly with long documents or links, Luis first guides employees to the relevant information. Each response follows a clear sequence: first, the employee receives a short and understandable answer to the question. Luis then states the concrete next step. Only afterwards does he provide additional details, policies or links. If important information is missing, Luis asks a targeted follow-up question instead of generating a very general response. If he cannot fully resolve the issue, he prepares the request so that it can be handed over to HR easily.

Goal of the nudge

The goal of the nudge is to shape the interaction with Luis so that employees reach a useful and reliable next step more quickly. Instead of merely providing information, Luis should give the most important answer first, clearly state the next action and only then offer additional sources. This helps employees interpret HR information faster and pursue their concerns independently. At the same time, unnecessary follow-up requests to HR should decrease without overlooking special cases or uncertainty.

    Needs analysis

    • Information overload
      Long answers, multiple links or numerous policies make it difficult for employees to identify the information that is relevant to them quickly
    • Missing next step
      Even when an answer is technically correct, it often remains unclear what the employee should do next.
    • Unclear context
      In individual HR cases, Luis may lack information such as role, country or the specific situation. Without a follow-up question, responses can easily become too general.

    Cause analysis

    • Verbosity Bias

      Language models tend to formulate extensive answers. As a result, the central message can get lost among additional information, explanations or sources.

    • Automation Bias

      Users may quickly trust a chatbot answer because it appears technical and objective. Without clear framing, general statements may be applied too strongly to an individual special case.

    • Over-Refusal bzw. Risikoaversion

      Out of caution, a bot may refer users to documents or HR in sensitive HR matters instead of first providing safe and clearly limited guidance.

Target Group

The target group consists of Lidl employees who use Luis for everyday HR questions. This includes both employees and managers, while the information displayed may vary depending on role and country. The HR department also benefits indirectly, as standard requests are solved or prepared more clearly and are less often passed on by email.

Added value of the nudge

  • Faster orientation

    A direct short answer and a clear next step help employees pursue their concern without lengthy searching.

  • Fewer misinterpretations

    Follow-up questions when context is missing, together with clear framing of limitations, reduce overly general or unsuitable conclusions.

  • Relief for HR

    Standard requests are solved or prepared more systematically, allowing HR to focus more on complex and individual cases.

CONTACT US

Lara Fricke

Veronika Gisbrecht

Larissa Lessing

Lilia Peters