<aside> <img src="i" alt="i" width="40px" />

Product requirements

Read PRD for product behavior, company policy and scope. Production readiness requires consistent incident data and a defined policy precedence model.

</aside>

AI agent UX · Product design · Self-directed concept · Alexis Ovitt

Relay explores how an AI agent can handle after-hours property-management operations while preserving human judgment, trust, and relationship work.

<aside> <img src="/icons/checkmark-line_gray.svg" alt="/icons/checkmark-line_gray.svg" width="40px" />

Evidence boundary

Relay is a concept prototype. Domain-informed findings, prototype-defined behavior, and unvalidated business hypotheses are labeled separately so seeded demo data is never presented as a real outcome.

</aside>

01 — Problem / Opportunity Brief

Tenant problem

During an emergency, a tenant is stressed and still has to decide what information matters.

Manager problem

After hours, the property manager becomes the coordination system: assessing urgency, finding vendors, authorizing work, and keeping the tenant informed.

Business risk

• Manager burnout / churn
• Tenant churn / renewal risk
• Reputation damage

<aside> →

Opportunity

Design an AI agent that handles repeatable emergency coordination while preserving human judgment where it matters.

</aside>

Tenant → Relay → Property manager

Relay handles the coordination. The property manager keeps the judgment, trust, and relationship.

02 — Domain Research Summary

Design trigger

A blank text field places too much cognitive burden on someone dealing with an active emergency.

Research input

I interviewed a water-restoration project manager to understand what responders actually need to know during a flooding incident.