Startups · AI

CX Conversational Designer

Flock Safety · Remote - USA · Remote

← All jobs
About Flock Safety

Backed by a16z, Kleiner Perkins and Y Combinator.

About the role

Not an unsupervised AI role. Trust-critical decisions, SEV incidents, and VIP/Safe City accounts stay human-reviewed. This role designs the guardrails; it doesn't replace the judgment call.

What they're looking for

  • Experience with Intercom/Fin specifically, or a comparable modern AI agent platform
  • Demonstrated ability to design and iterate conversational flows, including intent mapping, utterance training, and fallback strategies, in a live AI agent platform
  • Hands-on prompt development experience with LLM-based conversational agents, not legacy decision-tree bots
  • Fluency reading conversation logs and performance dashboards to diagnose failure points and prioritize fixes
  • Writing that reads naturally in a customer-facing bot voice: clear, concise, on-brand
More about this role

Flock's Fin/Intercom AI agent is the first response most customers get. That agent has to sound like Flock, resolve real issues, and know precisely when to stop and hand off to a human, especially on trust-critical accounts. Right now that design work is split across whoever has time for it. We need one person who owns the craft of conversation design as the agent portfolio grows past this first launch.

End-to-end conversation design for Flock's AI support agents, from intent mapping through prompt development to production deployment.

A measurable lift in AI-resolved conversations without a rise in bad handoffs or missed trust-critical escalations.

The escalation and fallback logic that routes SEV incidents, VIP accounts, and Safe City accounts to a human every time.

A conversation analytics practice that turns agent performance data into a prioritized backlog of design fixes.

Documentation and design standards that let the agent portfolio scale past one bot without design quality slipping.

Partnership with Support, Product, and Trust stakeholders to keep agent behavior aligned with SLA and safety requirements.

Not a scripting job. This isn't static decision trees; it's designing...

Read the full posting on Flock Safety's site ↗

Customer Experience

Build your edge while you search

Free tools for founders and investors, plus VC Unfiltered, our take on startups, venture and the people who build them.