The challenge
The firm’s deal team handled a constant flow of inbound NDAs and Confidential Information Memorandums (CIMs). Every NDA required manual review against the firm’s reference positions. Every CIM required an associate to extract and score data against the investment scorecard before the deal could be discussed. Both workflows were repetitive, high-volume, and pulled the team away from actual deal work.
What we built
Two AI automation pilots on shared cloud infrastructure, deployed in the client’s own environment:
- NDA redlining — a tool that monitors for inbound NDAs, evaluates them against the firm’s reference positions and redline history, and surfaces a structured deviation report to the deal team. The tool also recommends additional redlines outside of the firm’s playbook, catching provisions the team didn’t consider
- IC memo generation — a tool that monitors for inbound CIMs, extracts and scores relevant data against the firm’s investment scorecard, and produces a formatted one-page summary including market analysis, with model routing based on task type
How we did it
- 1
Discovery and infrastructure
mapped both workflows, delivered an infrastructure recommendation before any cloud contract was signed, then provisioned the environment (weeks 1-2)
- 2
Build and deploy
built both pilots in parallel on shared components, tested against real deal documents, validated with the deal team (weeks 3-8)
- 3
Managed service
Paradox Machines operates the platform on an ongoing basis — monitoring, pipeline health, and configuration updates as reference positions and scorecard criteria evolve
The role of AI
Claude powers both tools inside the firm’s environment. NDAs are checked against the firm’s actual negotiating history, not generic templates. CIMs are scored against the firm’s own investment criteria, so the one-pager reflects how this firm evaluates deals.
The results
- 10 hours saved per week across the deal team on NDA review and CIM screening
- Production-ready in 8 weeks — both tools live, documented, and operated under an ongoing managed service
Best fit
PE firms and deal teams drowning in repetitive document review who want AI tools running in their own environment, tuned to their own positions and criteria.