Generative AI makes it easy to produce a counter-proposal for any clause. The risk is that it produces a reasonable position rather than your position. Over hundreds of contracts, that drift adds up to real risk.
What a good playbook contains
For each clause type, a playbook sets out:
- Preferred: what you open with.
- Acceptable: what you will agree without escalation.
- Fallback: the furthest you will go, usually with approval.
Walkaway is different. It is not a position you offer. It is a scenario your team plans for, so it is modelled separately and kept internal.
How playbook-first works
When ContractSense reviews a counterparty’s paper, it compares each clause with your playbook. Where a clause falls outside your positions, the suggested counter comes from your playbook first. AI suggestions appear only after that, and they are labelled as such.
Reviews also combine two kinds of check. AI verdicts catch meaning: “this indemnity is broader than it looks”. Deterministic phrase rules catch the things that must never be missed: “there is no limitation of liability clause”. Together they are both smart and consistent.
Every change is a new version
Any AI-produced redline is saved as a new, immutable version with visible change tags. Nothing is overwritten. A person accepts or rejects each change, and the reason is recorded.
Playbooks also connect negotiation to enforcement. The positions you won become the terms ContractSense monitors after signature.
Want to know how your organisation compares? Take the maturity assessment.