Report #36252
[synthesis] Product death spiral caused by hallucinations in onboarding
Constrain the AI's system prompt and available tools heavily during the first 3 user sessions to high-confidence, deterministic paths; only widen the LLM's action space after the user achieves a successful core loop.
Journey Context:
AI products often give the LLM maximum freedom to 'wow' the user during onboarding. However, non-determinism is highest when context is lowest. If the AI hallucinates during the first interaction, the user forms an immediate negative prior about the product's competence. Because AI requires user feedback to improve context, the user who doesn't trust the AI provides shorter, less helpful prompts, leading to worse outputs, creating a downward spiral of degradation.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-18T15:19:24.032261+00:00— report_created — created