Worked example
Scope, and the clause behind every answer
Six change requests against one statement of work. Each one is answered with the clause it came from, priced where it is chargeable, and refused where the contract refuses it — and the variation allowance runs out in front of you while you read.
A consulting engagement rarely loses money in one decision. It loses money in six reasonable ones. The client asks for something sensible, somebody says yes because the relationship matters more than the paperwork, and the question of whether it was in scope gets settled months later in a conversation nobody wanted to have.
The answer is not a stricter firm. It is being able to say, the same day, which of three things a request is — already paid for, chargeable, or something the contract rules out — and to say it with the clause attached.
That last part is the whole thing. A change-control system that says “out of scope” without naming the clause is an opinion with software around it, and no client has ever accepted one.
What it checks
Already covered
Work the client has paid for. The failure here is taking two weeks to say so — or pricing it.
Explicitly excluded
The contract rules it out in so many words. The request is reasonable and the answer is still no.
Chargeable variation
Outside the agreed scope, so it is priced from the rate card and sent for signature.
The allowance
What is left of the variation budget. Most firms find this figure once it is already gone.
Already started
Someone said yes on a call. The hours are spent either way; what is open is who pays.
The end date
A separate question from the money, and usually noticed a month later than it should be.
Try it
Six requests, one statement of work
Synthetic change requests to Nilgiri Advisory Partners on SOW-2026-114. Pick any of them — the verdict comes with the clause that produced it, the rule that caught anything else, and the reply it would send. CR-05 and CR-06 are the two worth your time.
Change requests
5 needing a personVariation allowance, Clause 7.5
₹4,38,600.00 of ₹4,00,000.00
Over by ₹38,600.00 once all six are agreed
CR-01 · raised 2026-08-21
Already covered
Operations director, Harbourline
“Can you also look at how the rate tables handle multi-leg shipments? We think that is where the margin report goes wrong.”
- Fee
- No charge
- Completion
- 2026-11-27, unchanged
- Decision sits with
- Engagement manager
Because of this clause
Clause 2.1 of SOW-2026-114
“Review of the existing freight costing model, including assumptions, rate tables and the monthly margin report, and a written assessment of its accuracy.”
covers: costing-model, margin-report, rate-tables
What a person decides
The rules get you the position and the draft. Whether to hold a client to a clause is a judgement about the relationship, and the software does not make it.
No charge arises. The estimate below is what it would have cost if this had not already been covered — shown so the value of saying so quickly is visible.
- 12h Analyst₹33,600.00
The reply it would send
Subject: CR-01 — covered by the current scope
Thank you for CR-01, raised on 2026-08-21.
This is covered by Clause 2.1 of SOW-2026-114, which reads: "Review of the existing freight costing model, including assumptions, rate tables and the monthly margin report, and a written assessment of its accuracy." There is no charge and no change to the agreed dates.
Assembled from the assessment, not written by a model. It can only say what the rules above actually found.
Audit trail
What was asked, which clause answered it, and who agreed. This is the record that settles the argument at the final invoice — or prevents it.
- 2026-08-21Request receivedsystemCR-01 from Operations director, Harbourline, touching rate-tables, margin-report.
- 2026-08-21Verdict: in-scopesystemCited Clause 2.1 of SOW-2026-114.
- 2026-08-21No action neededsystemNo charge arises from this request.
Synthetic data. Nilgiri Advisory Partners, Harbourline Logistics Pvt Ltd, SOW-2026-114 and every figure in it are invented, including the rates — this is not a price list. The matching is rule-based rather than a model reading the contract: each clause declares what it covers, each request declares what it touches, and the overlap decides.
Being straight about it
What is real here and what is not
Real
- The clause matching, and the rule behind every verdict
- The pricing from the rate card, and the running allowance
- Approval routing, including the escalation when the allowance is breached
- The reply, assembled from the assessment rather than written
- The audit trail
Not real
- The firm, the client, the contract and every figure in it
- The rates, which are invented and are not a price list
- The reading of the request — each one declares what it touches, rather than a model inferring it from the prose
Why no model reads the contract
It could. We chose not to, and the reason is the output. This process produces a commercial position somebody may have to defend to a client, and a probabilistic answer that is right most of the time is the wrong shape for that. The matching here is tag-based and deterministic: each clause declares what it covers, each request declares what it touches, and the overlap decides. Where real judgement is needed — whether to hold a client to a clause, whether to waive a fee — it stops and asks a person, because that is a question about a relationship rather than about a contract.
All of the examples run on the same foundations. The money handling, the exception model, the approval routing and the audit trail are shared code — the rules for each process are its own. Invoices, quoting, candidate screening and database reactivation are the others.
Where does your scope actually go?
If change requests are settled from memory, or the variation conversation happens after the work is done, that is worth an hour.

