Your bank does not hold a clause. It holds 41 of them.
Treating a precedent bank as a template store throws away the only thing it really knows — not what the clause says, but where it landed across every deal, how far that is from where you opened, and who it moved for.
Five clauses, 41 deals each
Ordered strongest position to weakest. The bar is how many deals landed there; the tail is usually the interesting end.
Where 41 deals actually landed
- 100% of fees paid in the prior 12 monthsopens12
- 150% of fees paid9med
- 200% of fees paid7
- Capped, with carve-outs left uncapped8
- Uncapped5
Opens at
100% of fees paid in the prior 12 months
Lands at
150% of fees paid
Most common
100% of fees paid in the prior 12 months · 12 deals
The five uncapped deals are the same counterparty group, signed inside one quarter, all by the same partner. No individual file looks like a pattern. Forty-one of them do.
Drafted
The Supplier's total aggregate liability under this Agreement shall not exceed 150% of the fees paid in the twelve months preceding the claim, save for liability which cannot be limited at law.
Drafted at the median landing rather than the opening position, from the wording used in the nine agreements that closed there.
Termination for convenience is the healthy shape: the template opens at 90 days' notice and the median deal lands there too, which means the term is being accepted rather than conceded. 3 of the 5 clauses here do not do that.
5 of 41 agreements have no liability cap at all
Not one of those files looks alarming on its own. Each was a commercial judgement someone made on a Thursday with a deadline.
Counted together they are the same counterparty group, inside one quarter, signed by the same partner. That is a pattern, and it is invisible to anyone searching the bank one document at a time.
The part that argues against this feature
A system drafting from that bank will reproduce those five. It has no way to know they were a bad quarter rather than a house style, because the bank records what was signed and never what it cost.
Which is the whole reason the spread is shown and not just the output. A drafted clause that arrives without its distribution is asking you to trust your own worst deals.
A record, not a standard
Everything a precedent bank knows, it learned from deals you were under pressure to close. Four consequences worth holding onto.
- A bank is a record, not a standard
- It contains what you agreed under time pressure, with a client who wanted the deal signed, against counsel who was better prepared that week. Drafting from the median reproduces all of that faithfully.
- The tail is the interesting part
- Five uncapped liability deals in a bank of forty-one is not noise. It is a pattern with a name attached, and it only becomes visible when the deals are counted rather than searched.
- Volume is not authority
- A position appearing in eighteen agreements means it was accepted eighteen times, not that it is correct. If the firm has been opening in the wrong place for six years, the bank now says so with confidence.
- So it shows you the spread, not just the answer
- Every drafted clause arrives with the distribution behind it and the position it was taken from. A partner overruling the median is the system working, which is why the median is never presented on its own.
When the spread is not worth drawing
A distribution over eleven deals is an anecdote with axes on it.
- It drafts from what you have
- A firm with forty agreements has a thin distribution and a firm with four hundred has a good one. Below roughly thirty deals per clause type the spread is anecdote, and the page will say so rather than draw a chart.
- Clause type is not deal type
- A liability cap in a construction contract and in a software licence are different negotiations. Distributions are segmented by matter type, and thin segments stay thin rather than being pooled to look fuller.
- It cannot see what you walked away from
- The bank holds signed agreements. The deals that collapsed over a term, which are the most informative of all, left no document to count.
- Nothing here is a position on the law
- That a term appears in every agreement your firm has signed says nothing about whether it is enforceable. Drafting is assembly; enforceability is judgement.
The honest summary: this makes your firm's own history legible and drafts from it quickly. Whether that history is any good is a question it cannot answer, and the first useful thing most firms get from the spread is an argument about the template.
The rest of the platform
Assistant
Ask the file a question. Every answer cites its source.
Research
Nizam, regulations and ministerial decisions.
Arabic and English
Two versions that say the same thing.
Contract review
Marked against your playbook, not somebody else's.
Document sets
Thousands of pages, ranked by what matters.
Matters
Every document, deadline and hour against the file.
Time and billing
Captured as you work, not reconstructed on Thursday.
Deadlines
Counted from the trigger, not from memory.