Handoff-to-Human Architecture

The case

AI products that detect emotional/relational depth and route the user to a human (crisis line, pastor, friend, therapist).

Why it scores the way it does

Solution Fit 7 / 10

Adds friction in companion experience; direct UX shift

Stakeholder Fit 9 / 10

Innovators with chatbot products

Feasibility 6 / 10

Engineering and partnerships doable in a year

Urgency 7 / 10

Aligns with regulatory crisis-content requirements

Current landscape

Status: ● Nascent

988 crisis line routing for self-harm content exists in major chatbots. No general relational/parasocial depth detection or proactive handoff to humans.

References (4)

↗ Chatbots Need Guardrails to Prevent Delusions and Psychosis [Industry]
Argues for handoff to humans on crisis content
↗ Do You Feel Comfortable? Detecting Hidden Conversational Escalation in AI Chatbots [Academic]
Detection of emotional escalation - methodology for relational-depth detection
↗ 7 AI Safety Strategies for Therapy Chatbots [Industry]
Practitioner guide for crisis routing and escalation patterns
↗ Simulating Psychological Risks in Human-AI Interactions [Academic]
Research basis showing why handoff matters - dependency, addiction, psychosis simulations

Handoff-to-Human Architecture

The case

Why it scores the way it does

Current landscape

References (4)

Related