Loopkeeper

Loopkeeper is a B2B accountability intelligence platform built for multi-party processes - vendor -> partner -> client chains where dropped actions mean lost deals and damaged relationships In complex B2B workflows, accountability breaks down at handoffs. Emails go unread, escalations happen silently, and managers only find out when a client complains. No tool connects inbox signals, case health, and action ownership in one place. Loopkeeper deploys a coordinated crew of specialized AI agents that work together on every incoming email: 1. Analyst Agent - reads the email, extracts intent, urgency, and named parties 2. Router Agent - classifies the email to the correct case and assigns the next action owner 3. Action Agent - drafts a suggested reply or internal action item 4. Reviewer Agent - validates the proposed action before it surfaces to the user Beyond the inbox, a Coordinator Agent handles platform-wide routing decisions, an Escalation Analyst explains why a case went critical, and a Reporter Agent narrates manager digests. - Accountability Circle: visual map of who holds the ball right now - Pulse Score: 0–100 case health indicator updated by agent analysis - Reliability Scores: per-company and per-participant track records - Playbook Engine: rule-based automation triggered by agent-detected events - Audit Trail: full timestamped history with CSV/JSON export - Inbox Intelligence: 3-panel workspace with live Outlook integration and AI triage Next.js 15, React 19, TypeScript, Tailwind v4, AIMLAPI. Routing calls use Llama-4-Scout (lightweight); generation uses Qwen3.5; reasoning-heavy tasks use DeepSeek-R1. All agents share one config with deterministic fallback for offline operation. By making accountability visible and automating who-owes-what tracking, Loopkeeper lets B2B teams focus on relationships and cash collection - not chasing emails.