Solution · By problem
You're signing contracts nobody has read all the way through. The bill for that shows up two years later.
The big customer's framework agreement, the supplier's terms, the insurance policy, the SaaS that renews itself. It lands on Thursday, it has to be signed Friday, and somebody skims it. We automate the reading: the system checks every contract you receive against your own written position, flags the clause you should not be accepting, and watches the renewal dates nobody is watching today.
The problem
The risk is not in the contract you negotiate to death. It is in the twenty you sign without arguing.
- The contract lands on Thursday and gets signed on Friday. Somebody reads page one, the price and the term. The rest is signed on faith.
- There is no lawyer on staff — or there is, and they are the bottleneck: everyone waits for them to read, and they have forty things ahead of yours.
- The clauses that hurt —liability caps, late-delivery penalties, exclusivity, IP assignment, governing law— get accepted exactly as they arrived, because nobody had a written position to compare them against.
- Asking «what notice period do we have with this supplier?» costs half a morning, because signed contracts live spread across a Drive, an email thread and the desk drawer of somebody who no longer works here.
- The auto-renewal fires on its own. You find out when the invoice arrives for another year of a service you were going to cancel.
- When a contract goes back and forth three times, nobody rereads forty pages to find out what moved: you go by what was agreed over email and sign whichever PDF arrived last.
Cost of staying the same
Signing without reading does not get paid for on signing day. It gets paid on the day of the dispute, the day of the penalty, or the day of the renewal nobody saw coming. And it never shows up in a report under «contract review»: it shows up as a settlement negotiated down because the contract was not on your side, as twelve more months of a tool you stopped using, or as a missed deadline that was written down and nobody had read. To size the hole: the sector benchmark research puts the average value leakage from poor contract management at around 9% of contract value a year —the worst performers near 15%—, scope international. That is the size of the problem, not a promise from us.
The solution
A system that reads every contract you are sent, measures it against your position and keeps its eye on the dates
- 1The playbook comes first, because without it there is nothing to automate: which clauses you accept as they come, which you negotiate and which you never sign. Liability, payment terms, penalties, exclusivity, IP, data protection, jurisdiction. That document does not exist at almost any company, and it is what turns «this feels off» into a criterion a system can apply on its own.
- 2The system reads the incoming contract —PDF, Word or scanned— and takes it apart clause by clause: it works out what each one covers and checks it against your playbook. Out comes a short report: what sits inside your position, what deviates and in which direction, and what you should not be signing at all. It also flags what is missing: a clause that is not there is as much of a risk as a bad one.
- 3It pulls the contract record —parties, amount, term, notice period, renewal date, penalties, dated obligations— out of the PDF and into something you can actually query. The dates go into a calendar that warns you with room to decide, not the week it already renewed itself.
- 4When the contract goes back and forth, it compares versions: what the other side changed between v3 and v4, clause by clause, with the change marked and its impact spelled out. No more rereading forty pages to discover they moved a comma in the liability cap.
- 5The boundary is written down: the system flags and prioritises, a person decides and signs. A standard NDA inside your position moves fast; anything that deviates goes up to whoever needs to see it, with the reason and the quoted passage in front of them. And everything leaves a record: what was reviewed, against which version of the playbook, and what was decided.
What changes
What you stop losing
Value leakage attributed to poor contract management runs at around 9% of contract value a year, reaching 15% at the worst-performing organisations. Scope: international. It sizes the problem you are attacking; it is not a result we are promising you.
World Commerce & Contracting (WorldCC), «Poor Contract Management Continues To Costs Companies 9% Of Their Bottom Line», research report, 28 April 2020 — worldcc.com · scope: international
Review stops depending on who happens to be free on Thursday afternoon. The criteria are written down and get applied the same way to contract number one and contract number two hundred, in August, with the team on holiday.
Mechanism
Auto-renewals stop being a surprise: the notice date comes out of the contract the day it is signed and goes into a calendar that warns you while you can still do something about it.
Mechanism
What we measure: days from a contract arriving to it being reviewed, % of contracts with a complete record, deviations caught against the playbook, and renewals flagged with enough notice.
What we measure
Spec sheet
- Work it removes
- reading every incoming contract by hand to work out what you are accepting, then chasing the dates through the PDF afterwards
- Typical setup
- 3–5 weeks
- Input
- the contract a third party sends you: PDF, Word or scanned, by email or from the repository
- Output
- a deviation report against your playbook, the contract record (parties, amounts, terms, notice periods, obligations) and renewal and expiry alerts
- Works with
- SharePointGoogle DriveMicrosoft 365Google WorkspaceDocuSignSlack
- Can connect to
- Your clause playbookYour archive of signed contractsYour calendar of renewals and notice datesYour alerting channel (Slack, Teams, email)
- What we measure
- days from a contract arriving to it being reviewed% of contracts with a complete recorddeviations caught against the playbookrenewals flagged with enough notice
- Good fit for
- companies that sign third-party contracts often —suppliers, large customers, SaaS, leases, insurance— and have no in-house legal team that can keep up
- Not a fit for
- the one-off, high-value deal, where what you need is a lawyer negotiating, not a first-pass filter
Frequently asked questions
The direction the paper travels. Generating contracts is the outbound side: you close a deal and your document has to come out of your template, ready to sign. This is the inbound side: the other party wrote it, it arrives with their terms and their small print, and the job is understanding what you are accepting before you sign and not losing track of what it commits you to afterwards. One side speeds up the sale. The other protects the company from what it has already signed.
No, and anyone telling you otherwise is selling you a problem. The legal judgement stays with your legal team: the playbook is written by you with them, and it is theirs. What the system does is the mechanical work that eats their time today or, worse, that nobody does at all: reading the forty pages, finding where each clause is, comparing it against what you had already decided, and arriving with the deviations flagged and quoted. Your lawyer stops being a PDF reader and gets to decide on what actually needs judgement. And a person signs, always.
That is usually where the first value is, which is why we start with the archive. We sweep the back catalogue —what sits in Drive, in SharePoint, in email and in scans— and pull the record for every live contract: what you signed, with whom, until when, with what notice period and with which obligations outstanding. At almost every company that first sweep turns up two uncomfortable things: renewals already running that nobody consciously decided, and obligations you took on that nobody is meeting because nobody knew they were there.
You measure it before you trust it, and we show you the number even when it does not flatter us. Before letting it work alone we run it over contracts your team had already reviewed and compare: what it found, what it missed and what it over-flagged. Three things contain the failure. First: it does not answer «clean» by default —if a playbook clause does not appear in the contract, it reports that absence, because a missing clause is as much of a risk as a bad one—. Second: it quotes the exact passage it relies on, so checking its work takes seconds rather than another full read. Third: the line between what passes on its own and what goes to a person is yours to draw, by contract type and by value, and it moves when the data says it should.
It is the right question and it shapes the whole build, not an appendix at the end. It gets settled by writing down three things before we start: where the document is processed and under what contract with the model provider —including the commitment not to train on your data—, who inside your company can see what (the CEO contract is not open to everyone, and repository permissions are respected as they are), and what gets logged for every query. If your policy or your sector cannot accept a contract leaving your perimeter, we build it with a model inside your environment and trade some quality for not having that argument with anyone. What we do not do is build it first and ask afterwards.
Want it running in your business?
You’ve pinned the problem. We ship the fix and leave it measured.