LedgerIQ

Mission

AI “Decision Memory” for Teams Most companies already have: * Slack messages * Notion docs * Google Docs * Jira tickets * Zoom transcripts * Emails But nobody remembers: * Why a decision was made * What alternatives were rejected * Who approved it * What assumptions were true at the time Six months later, teams repeat debates, rebuild dead projects, or make contradictory decisions. The product An AI system that automatically creates a “decision graph” for a company. It: * Watches meetings/docs/chats * Detects decisions automatically * Extracts: * decision * rationale * risks * owner * expected outcome * dependencies * Builds a searchable memory layer Example queries: * “Why did we stop supporting Android tablets?” * “What assumptions led to the pricing change?” * “Which customer requests caused roadmap shifts?” * “Show all unresolved architectural decisions.” Why this is interesting now LLMs are finally good enough at: * summarization * entity linking * temporal reasoning * extracting intent from messy conversations But most AI startups are still: * generic copilots * wrappers * chatbots This attacks a painful operational problem with measurable ROI. Customers Start with: * 50–500 person tech companies * agencies * consulting firms * product orgs * remote-first teams Later expand into: * legal * healthcare administration * enterprise PMO teams * government contractors Business model Per-seat SaaS: * $20–80/user/month Or: * charge by indexed knowledge volume * enterprise compliance/security tier * private deployment Why companies would actually pay The ROI is tangible: * fewer repeated meetings * faster onboarding * reduced context loss * lower key-person dependency * preserved institutional memory If one senior engineer wastes 5 hours/week rediscovering context, the software pays for itself quickly. MVP You could build a strong MVP surprisingly fast: Inputs * Slack * Google Meet/Zoom transcripts * Notion * Jira Core workflow 1. Detect likely decisions 2. Ask users for lightweight confirmation 3. Store structured decision objects 4. Enable semantic search + timeline views Secret sauce The moat is not the LLM. The moat is: * accumulated organizational graph * integrations * trust * workflow embedding * historical context quality Competitive edge opportunities Most existing “AI knowledge tools” are passive search. This becomes: * organizational intelligence infrastructure Potential advanced features: * contradiction detection * “this decision conflicts with previous policy” * assumption expiry alerts * decision outcome scoring * predictive warnings (“similar launches failed before”) Distribution strategy Strong channels: * LinkedIn founder content * engineering leadership communities * productivity YouTube * Slack/Notion marketplaces * integration partnerships A killer demo would spread naturally: “Ask your company why anything exists.”