Where you already outsource is where AI lands first
There is already a ready-made opportunity map inside your company, and it is not in the technology plan. It is in the contracts you already pay every month.
Look at the contracts your company already has with suppliers as a list of priorities: see how much of each contract is work that follows a rule, know that the budget is already approved, and choose which contract to review first.
Sign up once to unlock this course
Course 01 is open. For courses 02 to 05 we ask for six fields. It is a single signup: done once, valid for the whole track. It is not a free diagnosis and it does not trigger automatic sales contact.
The answer to “where do we start” is already written down somewhere
When the board asks where to apply AI first, the conversation turns inward: which processes hurt most, which team complains most, where the spreadsheets pile up. That search is expensive, because everything in-house looks like a candidate and nothing has a declared price.
There is a shorter path. Every company already pays outsiders to do work that requires expertise, and pays with a contract, an invoice and a cost centre. Examples: the law firm reviewing standard contracts; the company checking documents; the agency producing variations of the same creative; the contact centre answering the same thirty questions; the consultancy assembling a monthly report from data you send them.
Those contracts have three things no internal process gives you for free: a known price, a written scope and a measured volume.
A supplier contract is an opportunity with the price on the label
An internal AI project has to fight for new budget. Reviewing a supplier contract touches money that already left the account last month. That changes who has to approve it and how long the approval takes.
And the contract gives you for free what an internal project takes weeks to assemble: a description of the work. It is in the scope, the agreed turnaround times and the price-per-volume table. Someone was already obliged to write down what has to be done, in how much time, and what counts as done well.
Volume per month, price per unit, agreed turnaround, how much error is tolerated, what needs your approval, and how much of the work follows a written rule versus how much depends on the judgement of whoever performs it.
How much of the contract follows a written rule
No contract is 100% automatable, and presenting it as if it were is the fastest way to lose credibility in the first meeting. What exists is a proportion. In a document-checking contract, almost all of it can follow a rule, because the criteria sit in a public regulation anyone can read. In legal counsel negotiating a delicate clause, the proportion is small, because the value is in judgement under pressure.
The estimate does not have to be exact. It has to be defensible. Take the number of cases in a month, separate the ones that follow the normal path from the ones that need an exception, and you have the proportion. One warning: a contract with 80% standard cases does not turn into an 80% saving. It turns into an arrangement where the 20% of exceptions finally get attention from experienced people.
Starting inside usually stalls, and the reason is not technical
Automating an internal process means touching a team that has a manager, a target and a history. The gain shows up as “people with more free time”, which nobody knows how to turn into a number. The cost shows up as resistance, which everyone feels the same week.
Reviewing a supplier contract flips that around. The gain shows up as a cost line going down, and finance knows how to measure that without argument. And no internal team is defending territory, because the work sits outside the company.
This is not an argument to outsource more, nor to cut suppliers for sport. It is simply the observation that the first change has to be the one whose result is visible without argument — because that is the one that will fund the second.
When the contract is not the way in
Three cases where it does not work:
Low volume. If the contract has little volume, the effort of building the solution does not pay for itself.
What you are buying is liability. If the real value of the contract is having an outsider carry the legal risk, bringing the work in-house means bringing the risk in too. That has to be discussed first.
The supplier holds the data, you do not. Then the negotiation comes before any solution design.
There is a fourth case too: the supplier already uses AI and passes part of the gain through in the price. There the opportunity is not to replace them. It is to renegotiate with information, knowing which part of the service stopped being done by people.
Basis
This course does not propose cutting suppliers. It proposes using the contract as a source of information: it already contains volume, price, turnaround and criteria — exactly what an internal idea born of intuition lacks.
When the inventory becomes an investment decision
If the first contract on your list has meaningful volume and more than 60% of work that follows a rule, it is worth assessing. Prumo Discovery is TheNeil’s method for turning that candidate into a decision: it measures the current situation, assesses whether it is feasible and presents a recommendation.