A coaching chatbot people weren’t returning to.
The company coached middle managers through human coaching sessions and its own learning platform. An AI component could extend that support into the manager’s everyday work.
The central problem was engagement. Participants were not returning to the chatbot often enough. Responses that felt unhelpful, repetitive, or disconnected from the person’s situation could end the conversation.
The assistant needed to understand the coaching context and offer a useful next step without making the participant do too much work.
A clearer design for a coaching companion.
The work produced product concepts and an architecture for a more contextual Slack companion: one that could connect learning and coaching activity, offer concise prompts, and accept different forms of feedback.
The engagement centered on discovery and architecture. The weekly prompt and several interaction mechanisms were explored as design concepts; the outcome described here is the proposed product approach.