The AI savings in your deal model are a guess.
We test whether they are real, and we check how AI is already used inside the target, before you set the price.
The AI Operating Map is a task by task record of how a business really works: who does each task, who owns it, and where AI already operates. Every saving we report ties back to it.
Top down savings numbers fall over after close.
Most deal models carry an AI saving. It comes from a benchmark, or a percentage taken off a cost line. The number looks tidy. Nobody has checked it against how the work actually gets done in that business.
Then the deal closes. The first operating review asks which tasks the saving comes from and who owns it. Nobody can say. The saving slips a year, shrinks, and drops out of the plan.
A benchmark is an average of other companies
It says nothing about this target's systems, data or people. The investment committee is being asked to price someone else's results.
Savings sit on a cost line, work sits in tasks
A percentage applied to a function ignores which tasks can change, which cannot, and what has to be true first.
Nobody owns the number
Without a named owner for each line, the saving belongs to the deal model and to no one in the business.
Industry Reference Maps from day one.
For each industry we hold a task level reference model of a typical business*. It shows which roles and tasks AI can change, and what that change is worth. We do not start from a blank page or an industry average. You get a first view in days of where the savings are likely to sit.
Aim first
We take the saving in your deal model and compare it with the reference business. That shows quickly whether it is plausible and worth pursuing.
Then work backward
If it holds up, we trace it to the roles and tasks that would have to change. Each saving has a name and a place in the business.
Then test it in the target
The reference map is the hypothesis. We check it against how the target really works and report what holds, what does not, and the confidence on each line.
* Industry Reference Maps are built from public data and our own research. No client data is used to build them.
We map how the work really gets done, task by task.
Every saving we report traces back to a task, a person who does it today, and the conditions that have to hold for it to change.
Map the work
With the people who do it, we list the tasks in the key functions, how long they take and where the hand offs are. This is the ground the numbers stand on.
Test each saving
Every line is sorted into what is real now, what is possible with change, and what is hope. Each carries a range and a confidence level you can show a committee.
Test the capacity to deliver
We score whether leadership, data and processes can carry the change. A saving the business cannot deliver is a risk to the price.
What AI is already doing in the business you are buying.
The target is almost never starting from zero. Staff use AI tools, some approved and some not. That changes the numbers and it changes the risk.
Savings already captured cannot be counted twice
If the target already gets a benefit from AI, it is in today's cost base. Counting it again in the model overstates the upside.
Unmanaged AI use is inherited risk
We look at tools in use, unofficial use, contracts, and what company data is going to outside tools. The buyer takes on whatever is there.
The AI strategy slide gets tested
Management will have a view on AI. We compare it with how the work really runs and report where the two differ.
One page, built for an investment committee.
This is a fictional example. It shows the shape of what you receive. Figures are illustrative only.
Example Co, a mid market services business
| Function | Deal model says | Tested range | Confidence | Condition |
|---|---|---|---|---|
| Customer support | $2.4m | $0.9m to $1.5m | MEDIUM | Knowledge base cleaned up first |
| Finance operations | $1.1m | $0.6m to $0.9m | HIGH | Two systems already connected |
| Sales operations | $0.8m | $0.1m to $0.5m | LOW | Customer data is in four places |
| Claims handling | $1.6m | $0.0m to $0.4m | LOW | Process is not written down |
AI already in use
- Support team uses an approved chat assistant. Its benefit is already in the cost base.
- Finance staff use two unapproved tools. Customer data is pasted into one of them.
- The strategy slide says AI is "embedded". We found it in two of six functions.
- One vendor contract allows the vendor to train on the target's data.
Readiness
Flags for specialists
- Data going to outside tools: privacy review recommended.
- Vendor data terms: legal review recommended.
Fixed fee. Fixed scope. Clear timeline.
All prices are in AUD plus GST.
NDA first. Then a short, defined list of access.
What we need from the target
- Time with the heads of the key functions
- Process documents and system lists, where they exist
- The list of software and AI tools in use, with contracts for the main ones
- The management presentation on AI and the cost base by function
What we do not need
- Source code or direct access to production systems
- Personal data about customers or staff
- More than a few hours of any one person's time
Week 1
NDA signed. Access agreed. Task mapping starts in the key functions. First read on AI use inside the target.
Week 2
Interviews and document review. Each saving is tested and given a range and a confidence level. Contracts and data flows reviewed for flags.
Week 3
Readiness score. Strategy tested against reality. Recommendation and a summary written for the investment committee, with a working session with your team.
Delivered by Timothy J. Hitchens, Managing Partner.
The plan carries on into the business.
The 100 Day Plan hands the tested savings to named owners. From there, the AI Framing Sprint sets how the business will lead AI, and the AI Operating Map tracks each task, owner and result over time.
What deal teams ask first.
How do you handle confidentiality?
An NDA is signed before we see anything. Findings go to you and no one else. We work in line with the confidentiality terms of the deal, including any clean team rules you set.
What access do you need to the target?
Time with the heads of the key functions, process documents and system lists where they exist, the list of AI and software tools in use, and the management presentation. We do not need source code, production access or personal data.
What if the target will not give access?
We work with what is available and say plainly what we could not test. The AI Value Screen runs on the information you already have. A target that will not allow access to basic information about how work is done is itself a finding for the committee.
How do you stay independent?
Fees are fixed and agreed before we start. We are never paid on the size of the savings, so we have no reason to make them look larger or smaller. We sell no software and take no referral fees from vendors.
How is this different from the big firms?
You work directly with the person who built the method, not a rotating team. The scope is narrow: whether the AI savings are real and how AI is used today. The fee and the timeline are fixed.
What do you not cover?
We do not perform legal, security or privacy reviews. We flag the risks we find and recommend you bring in specialists. We do not value the business, and we do not write the deal model.
Book a 20 minute deal call.
Tell us about the deal and the savings in the model. We will say whether a screen, a full diligence or neither makes sense.
Book a deal call →