Operating Memory Lead

About the role

You design and own the operating memory of a fast growing AI native commercial insurance company. You turn scattered human judgment into structured, retrievable knowledge that AI agents can safely reference at scale. You embed directly with operators across sales, intake, service, and underwriting to capture how work actually happens, not how leaders assume it happens. You translate transcripts, calls, and Slack debates into source of truth artifacts that compound over time. You ensure that every process, edge case, and rework is documented in a way that both humans and machines can understand. You make meetings produce durable artifacts instead of forgotten notes. You build the playbook layer that allows Harper to scale without losing the nuance that made it work.

Key facts

What you'll do

• Capture tribal knowledge by embedding with sales, intake, service, placements, and renewals, sitting with operators, shadowing workflows, listening to calls, and reading transcripts to document what people "just know."
• Build operating memory by turning transcripts, Slack threads, Looms, and one off explanations into source of truth docs, decision logs, playbooks, process maps, onboarding paths, and glossaries that persist beyond any single person.
• Use AI as a force multiplier by designing repeatable workflows that convert raw context into decisions, owners, open loops, SOPs, training material, and product requirements instead of relying on occasional summarization.
• Make meetings AI legible by shaping conversations in real time so they produce clear artifacts, defining decisions, owners, definitions, edge cases, unresolved questions, and next steps that can be retrieved later.
• Find the edge cases where workflows break, documenting reworks, escalations, stale quotes, underwriter follow ups, payment or binder gaps, COI delays, and customer confusion before they become systemic failure modes.
• Translate ops into product by sitting between operators and engineering, capturing what people actually do, where tools fail, what workarounds exist, and what needs to be built to make knowledge machine readable.
• Maintain the knowledge base by keeping docs current, assigning owners, killing stale guidance, and ensuring people know where the truth lives so that new hires and AI agents can find the same answer.
• Turn repeated problems into systems by recognizing patterns that appear three times or more and converting them into playbooks, QA checks, training artifacts, or product requirements that prevent future rework.
• Partner with leadership to prioritize which operating memory investments will unlock the most leverage across sales, service, underwriting, and compliance as the company scales.
• Define and track knowledge health metrics, such as reduction in repeated questions, faster onboarding, fewer escalations, and higher AI agent success rates, to prove the value of structured operating memory.
• Run lightweight discovery sprints that compress weeks of tribal insight into days of synthesized artifacts, enabling rapid alignment between operators, builders, and executives.
• Coach operators on simple documentation habits that make their work AI legible, reducing ambiguity for both humans and downstream systems that depend on clear context.

Requirements

You have 2-8 years of experience in research, product ops, knowledge management, technical writing, implementation, chief of staff work, qualitative research, instructional design, or startup operations, and you have shipped tangible artifacts that changed how teams work. You practice exceptional written communication and can turn a messy transcript into a clear operating document the same day without losing nuance or edge cases. You are genuinely AI native in practice, with a trained eye for when an output is structurally wrong, not just stylistically off, and you prompt for extraction of decisions, contradictions, owners, and edge cases rather than for simple summarization. You are intensely curious about how organizations actually work, and you like spending time with operators in their natural workflows instead of relying on slides alone. You notice hidden assumptions, missing ownership, and contradictions that other people walk past in meetings and documents. You prefer structured thinking but not bureaucracy, and you care more about whether documentation changes behavior than whether it looks polished in a slide deck. You are low ego, persistently curious, and allergic to the phrase someone should probably document that, preferring to clarify, capture, and codify in the moment. You are based in San Francisco or are willing to relocate there to work full time. You are comfortable in a fast moving, ambiguous startup environment where processes are invented on the fly and iterated on rapidly. You have demonstrated ability to interview stakeholders and extract operational detail without imposing your own framework on how people actually do their work. Your background may come from qualitative or academic research, ethnography, instructional or curriculum design, knowledge management, product operations, technical writing, research operations, implementation, chief of staff, library and information science, AI ops or human in the loop work, or messy startup operations, and the exact background matters less than the combination of writing rigor, AI tool fluency, structured information architecture instincts, and a bias for direct observation of real work.