Designing ResolveX to say "I don't know"
The worst thing a website agent can do is answer confidently and wrongly. ResolveX treats the knowledge boundary and a graceful human handoff as core product behavior.
Website chatbots fail in a predictable way: asked something outside their knowledge, they improvise. For a support widget that is embarrassing; for a product answering pricing or compliance questions, it is a liability. ResolveX treats refusal as a first-class outcome, not an error state.
Evidence-gated answers
ResolveX answers from approved knowledge and keeps its source trail visible. If evidence is thin, contradictory, or absent, the honest move is a handoff: 'I don't have a reliable answer; here's a human who does,' with the conversation context attached so the visitor does not repeat themselves.
- Retrieval confidence thresholds gate generation — no passage, no answer
- Citations are rendered in the widget, so visitors can verify instead of trust
- Handoff routes carry the conversation, source trail, and attempted steps into the shared inbox
- Workspace owners can review automated decisions and tune what the AI is allowed to use
The same evidence gate is applied before an answer is shown, while confidence rules decide when a person should take over. That makes the boundary inspectable instead of relying on the model to describe its own certainty.
ResolveX
LiveAI customer support that learns your business, resolves routine questions, and hands the hard ones to people with context intact.