Two questions. Answered separately.
Does this clause match our position? And would it actually hold up in court? Every other tool answers one of those and calls it risk. Vern answers both, independently, in a single pass, and the cases where they disagree are the ones worth your afternoon.
Does this match our position?
Measured against your playbook and nothing else. A number out of a hundred and a band, Accept, Negotiate or Escalate, so it's obvious whether this is a signature, a conversation or a phone call to a solicitor.
Would it hold up in the real world?
Enforceability and commercial viability under UK contract law, judged without reference to your playbook at all. A clause you're happy with can still be worthless, and that is not something your own positions can tell you.
One score gives you three answers. Two give you nine.
A single risk rating collapses two different problems into one colour. Here is every combination, with a real clause in each, the diagonal is where the two agree, and the corners are where the money is.
Liability for indirect and consequential losses is uncapped, with an exclusion that purports to remove all liability for such losses without limitation as to amount.
Your playbook requires a cap at two times annual fees, so this breaches it outright. Separately, a blanket exclusion with no cap is the kind of term a court reads against whoever drafted it, so it protects nobody.
Escalate, then send W1: a mutual cap at twice the fees paid in the preceding twelve months. Unlimited exposure on a £48k engagement, sitting outside your £2m aggregate cover.
Two of those nine cells are why we built this.
A single risk rating cannot express either of them. Both are common, and both cost real money.
Protection that isn't protection
The clause does exactly what you asked for, so your playbook waves it through, and it would collapse the first time you tried to rely on it. Perpetual confidentiality, fifty-mile covenants, penalty-shaped liquidated damages.
You only find out when you need it, which is the worst possible moment. This is the cell a playbook-only tool is structurally incapable of seeing.
Nothing legally wrong with it
Well drafted, enforceable, and completely one-sided. A generic legal-AI tool has nothing to flag because nothing about it is unusual, it's simply a good clause for them.
Only your own positions make this a problem. This is the cell an enforceability-only tool cannot see.
One pass. Two answers. Not two reviews.
Both scores come out of the same scoring stage, per clause. Running the contract twice would double the cost and let the two answers drift apart; producing them together keeps them consistent and lets Vern reason about the pair, which is what the cross-clause pass then builds on.
Scores come back through an enforced schema rather than text we hope to parse, so a malformed answer fails loudly instead of quietly becoming a wrong number.
Filled is yours. Outlined is the law.
The two pills look different everywhere they appear, so you never have to ask which number you're reading. It doesn't vary by screen, by density, or by anyone's preference.
Find out which of your clauses are in the corners.
We'll score an anonymised contract from your sector. The interesting part is never the clause you were worried about.