Harvey CX Knowledge Specialist at Harvey owns and consolidates product knowledge into a single foundation for both people and AI. The role involves auditing content, maintaining knowledge accuracy, and supporting AI Support agent effectiveness.
Responsibilities
Harvey ships product every week, and every release changes what our customers, our support specialists, and our AI Support agent need to know. Today that knowledge lives across fragmented surfaces, and the gap between what's shipped and what's documented shows up directly in ticket volume, resolution time, and the quality of answers customers get.
The CX Knowledge Specialist owns the source of truth: the articles, troubleshooting guides, and product knowledge that specialists and our AI Support agent answer from. You'll consolidate that knowledge into one well-structured foundation built for both people and AI, keep it accurate as the product moves, and use ticket, search, and agent data to decide what to fix next.
You'll work alongside our Enablement Specialists, who turn that knowledge into training and get the team ready for each launch. This is a hands-on role for a strong writer and systems thinker who wants to build the knowledge layer an AI-first support organization runs on.
Audit today's fragmented content surfaces: inventory and tag what exists, flag what's duplicate or stale, and migrate it into one knowledge foundation.
Own the knowledge base: product support articles, troubleshooting guides, how-tos, and product best practices that specialists and the AI Support agent rely on.
Structure knowledge for both humans and AI. Apply taxonomy, metadata, and templates consistently, and write articles that retrieve and chunk cleanly for the AI Support agent as well as for specialists.
Keep knowledge current through every release. Track what Product ships, and update or create the affected articles before launch so they're ready for the Enablement Specialists' training and for the AI Support agent.
Close knowledge gaps at the source by owning the feedback loop. Triage reports of inaccurate, missing, or buried content from downstream teams, then update the source material.
Mine ticket signals, search logs, and agent escalations to find missing content, stale articles, and areas with high deflection potential, then prioritize fixes.
Build AI-assisted workflows for drafting, refreshing, and detecting stale content (prompts, automations, review queues) so knowledge scales without adding headcount.
Run review cycles, ownership assignments, versioning, and archiving so content stays accurate as the product changes weekly.
Serve as the day-to-day contact for SMEs in Product, Engineering, AI Support, and GTM Enablement to get content reviewed, validated, and adopted.
