Harper Ops on Autopilot
Inspired by Harper — AI-native commercial insurance brokerage. Loop until the workflow is current, exceptions are owned, and human sign-off is captured where required.
Inspired by Harper
by Trooper
/loop 30m Start the "Harper Ops on Autopilot" loop. Inspired by Harper (https://www.harperinsure.com). Goal: open work triaged, exceptions owned, and core finance workflow current with audit trail Max iterations: 20 Between iterations run: Report open queue items, stale tasks, failed automations, and items awaiting human approval for Harper Exit when: zero open items without owner or explicit escalation, all external actions approved or sent, and systems of record current Step 1 — Intake requests: Pull new claims, policies, or quote requests from email, portal, and integrations. Step 2 — Gather evidence: Collect photos, estimates, docs, and third-party data required for adjudication. Step 3 — Evaluate and negotiate: Apply policy rules, compare to benchmarks, and draft resolution or counter. Step 4 — Human approval: Escalate exceptions and send customer-facing decisions only after sign-off. Step 5 — Resolve and report: Close loop with audit trail, update carrier systems, and flag renewals. ## Before you start Connect plugins: - GitHub (required) — Read branches, PRs, reviews, checks, workflow runs, and source diffs. - Browser / Web access (required) — Open pages, inspect live state, collect evidence, and verify changes. - Gmail (required) — Read incoming mail and prepare safe draft replies for approval. - Slack (required) — Post summaries, approvals, blockers, and handoff updates. - Support tickets (required) — Read customer tickets, feedback, onboarding gaps, and support themes. - CRM (required) — Read accounts, deals, owners, stages, and onboarding state. - Linear / Jira (required) — Read roadmap, task, bug, owner, and status context. - Notion (required) — Read and update approved briefs, docs, calendars, and reports. Attach skills: - Loop runner (required) — Self-pace iterations, run the check between passes, and stop only on the exit condition. - Code change + local verification (optional) — Edit code safely, run commands, and keep changes scoped. - CI debugging (optional) — Read failing checks, logs, and the smallest actionable root cause. - Approval workflows (optional) — Keep outbound actions in draft or approval states when risk is non-trivial. - Operations briefing (optional) — Summarize sources into concise, owner-aware, non-invented updates. - Customer operations (optional) — Identify at-risk accounts, missing setup steps, and safe next actions. - Documentation QA (optional) — Check docs against current product behavior and keep claims accurate. - Content operations (optional) — Turn signals into reviewable briefs while preserving source attribution. - Security triage (optional) — Assess alerts, prioritize patches, and avoid unsafe shortcuts. Self-pace this loop. After each iteration, run the check command, read the output, and only continue if the exit condition is not met. Stop when the exit condition passes or max iterations is reached. Give a short status update each pass.
Paste the kickoff prompt into Cursor, Claude Code, or Codex. Deeplinks do not install hook files.
1. Intake requests
Pull new claims, policies, or quote requests from email, portal, and integrations.
2. Gather evidence
Collect photos, estimates, docs, and third-party data required for adjudication.
3. Evaluate and negotiate
Apply policy rules, compare to benchmarks, and draft resolution or counter.
4. Human approval
Escalate exceptions and send customer-facing decisions only after sign-off.
5. Resolve and report
Close loop with audit trail, update carrier systems, and flag renewals.
Guardrails
Rules the agent must follow so it cannot cheat the exit condition.
- Require human approval before customer-facing sends, payments, or legal submissions unless pre-approved templates apply
- Preserve full audit trail linking source data to every automated action
- Escalate compliance, safety, or regulatory-sensitive items immediately
- Do not post ledger entries or bind coverage without explicit human approval
More Finance loops
Inbox Triage with Approval
On an interval, classify incoming mail, draft safe replies for routine threads, and escalate anything that needs a human decision.
Morning Operator Brief
Daily interval loop that reads your calendar, open tickets, and inbox priorities, then delivers a concise operator brief with ranked actions.
Customer Onboarding Watch
Interval loop that watches new signups, runs the onboarding checklist against each account, and nudges or escalates stuck users.
