“Oh, You’re Right” Is Not a Review

Have you ever run a 20-page contract through AI and watched it miss the clause that actually mattered? You point it out and the model replies, “Oh, you’re right.”

That is not a one-off. That is a structural problem, and it gets worse as the stakes get higher.

AI practitioners are under pressure to deploy AI on contracts, policies, code reviews, and other workflows without increasing risk. Yet most teams still rely on chat-style prompts and generic tools that miss key clauses, skip steps, and make it hard to prove what was actually reviewed.

Join Calvin Hendryx-Parker (Six Feet Up CTO and AWS Hero) and Whit Morriss (Six Feet Up Sr. Architect) for a live walkthrough of an AI-driven contract review system built entirely with open source Python tooling — SQLite, ChromaDB, lightweight agent skill definitions, and specialized subagents that route, validate, and escalate findings.

You’ll see how to turn a vague “review this” request into a structured series of sub-jobs with clause-level analysis, traceability, and audit-ready output.  

Walk away knowing how to:

• Chunk documents and tasks so each sub-task has clear acceptance criteria and nothing is silently skipped
• Manage context with structured memory instead of overstuffed prompt windows
• Build specialized subagents that route, validate, and escalate findings
• Prove what was reviewed, when, and by which agent
• Apply the same agentic patterns to Python engineering workflows like code review and test planning

No chat-style demos. Just a live Python-based system doing real work.