Director of Customer Service, Skylight

• 5+ years in CS/ Operations support, CS/Operations program management, with 5+ years of project implementation and major cross-functional programs
• You've owned, scaled and optimized an existing program leveraging AI and automation tools. Scaling and seeing a program to maturity matters more here than just the 0→1 experience
• You've lead voice, chat, and email support channels
• Hands-on fluency with LLM-based tools: prompt iteration, knowledge hygiene, evaluation of output quality. You don't need to code, but "I just used the vendor defaults" is disqualifying
• Deep metrics fluency: you can explain why deflection rate alone is a vanity metric and what you'd pair it with
• Proven frontline change management — you've brought skeptical teams through a tooling transition and they'd say good things about how you did it
• Structured vendor evaluation and management experience
• Strong written communication; you'll be writing the operating procedures, business cases, and exec updates
• You are comfortable operating independently, spotting ambiguity and proposing solutions

Desirable

• This role owns AI across our entire Customer Service operation, on two fronts with equal weight:
• Customer-facing AI — Own and strengthen our existing live AI agents across email, chat, and voice: raising resolution quality, expanding what they can safely handle, and carrying 24/7 coverage through seasonal spikes
• Agent-enablement AI — internal tooling that makes our human team faster and better: agent-assist and drafting, AI guided troubleshooting, QA automation, knowledge management, smart routing, summarization, insights mining. Determine if we build tools in house or vet and hire third party vendors
• We are explicit about our philosophy: we optimize for resolution, not deflection. AI at Skylight exists to remove robotic work from humans and to make the human moments better, especially for customers who need patience and care
• Scale the Program: Own the roadmap to scale support capacity for 2x company growth without linear headcount. Make smart build-vs-buy decisions and hold vendors (Decagon, Sierra, Intercom Fin) to live performance, not sales demos
• Drive Live Performance: Own resolution, containment, and CSAT across chat, email, and voice. Run root-cause analysis when AI fails a customer and ship the fix
• Manage Knowledge Infrastructure: Keep retrieval sharp by mining ticket logs for knowledge gaps, maintaining taxonomy, and ensuring Day-1 AI-ready content for all new product launches
• Build Frontline Trust: Partner with human support teams to roll out agentic tools, create feedback loops, and write the SOPs governing human-AI workflows
• Enforce QA & Engineering Discipline: Treat prompts and knowledge updates like production code using test suites, "break the bot" pre-releases, and staged rollouts
• Measure Business Impact: Honestly report ROI, cost per contact, and capacity returned against hiring alternatives—validating performance through live testing
• Lead as Player-Coach: Execute hands-on alongside our current AI support team while laying the foundation to scale the function