02How we scope
How we scope an engagement: case by case, agreed before anything is built
No two businesses hand over the same work, so no two engagements are scoped the same way. Each one is agreed case by case, on a call, and written down before the team builds anything. This page is the full mechanics behind the AI Praxis puts into your business: what the call establishes, what the diagnostic produces, how the measure and its starting point get fixed, how the result gets checked afterwards, and what happens if it turns out to move nothing at all.
01
Scoped on the call, not on a page
The commercial side of an engagement is agreed case by case, on the call, with the business in front of us. It is not published here, because a structure printed on a web page is a structure chosen before anyone has seen the work, and the work is the only honest basis for scoping it.
What the call does establish is whether there is anything here worth building at all. You describe how the business runs and where the day gets eaten. The team tells you which of that work AI can take over, which of it cannot be taken over yet, and what it would build first. If the answer is that there is nothing here worth building, we say that on the call.
The earliest thing this costs you is your own time and the access we need to see how the work actually happens. Everything else about the engagement, its shape, its length and its terms, is settled in writing before the build starts, and it is settled for that business and that work, not copied from a template.
02
What the diagnostic produces
If the call goes somewhere, the next step is a two-week diagnostic, scoped on the call. The team spends those two weeks inside the real work: sitting with the people doing the jobs, counting how long things take, and reading what your systems already hold.
At the end you get a written map of every job AI could do in the business, ranked by what it is worth and how hard it would be to change. It also names the jobs that should stay with people. That document is yours whether or not you go further, and it is the thing the scope is built from.
For a rough sense of the hours involved in your own business before any of that, the calculator does the arithmetic in front of you. The published ranges behind it, and why most of the figures circulating online do not hold up to checking, are catalogued on where every number comes from.
03
What incremental means
Everything the team builds is measured as incremental: new against the old way of doing the same work, not the whole of what the business earns. The question is always the same: what changed here that would not have happened otherwise.
Some of that change is money saved: fewer hours paid for the same output, fewer mistakes redone, fewer late charges paid because something sat too long. Some of it is money that was already available and was being dropped: a lead that went unanswered on a Saturday, a renewal nobody chased, a request that quietly went to a competitor because no one replied in time. Both count the same way against the starting point.
04
How the measure is agreed
This is the part most people want spelled out in full, so here it is.
Before anything is built, the business and Praxis agree a measure in writing. It is usually built from three things: what the work costs today, how long it takes today, and how much of it actually gets finished today. Your finance people choose which fits the work being automated. Neither side picks the measure that flatters the eventual result, because it is fixed before either side knows what the after number will look like.
Every measure comes from records the business already keeps: time sheets, ticket systems, invoicing software, not a new system built to manufacture a favorable starting point. Once agreed, the measure and its source are written into the same document that sets out the scope, and they do not move after that. That document is signed by whoever owns the finance function, usually a controller, a finance lead or the owner, and by Praxis. After the build, the identical measure is read again from the identical source, by the identical process.
If a business does not already keep the records a given measure would need, the honest answer is that a starting point for that work cannot be reconstructed from history that was never recorded. Tracking then starts from an agreed date before the build begins, and the starting point is whatever that new, simple tracking shows over an agreed period, not a number either side guesses at. It is slower, and it is the only version of this we consider honest.
Finance is usually the clearest place to see this working end to end, because the records already exist for reasons unconnected to us. How this plays out inside a finance team is written up separately.
05
How it is measured afterwards
The same measure, from the same source, is read again on a cadence set out in the same written agreement that fixed the starting point, not decided later by whoever finds it convenient. Whoever read the starting point reads it again: usually the business's own finance function, sometimes a bookkeeper or accountant the business already uses for other work. Praxis does not read its own number. A measure the team both proposes and grades is not a measure, it is an opinion, and this arrangement does not run on our opinion of our own work.
If a disagreement comes up about what the after number shows, it gets resolved by returning to the written definition agreed at the start, not by negotiation between the two sides. Whoever reads the records reads them the same way both times, independent of anyone with a stake in the answer. A measure that can be argued with after the fact was defined too loosely at the start, which is why that work happens before the build, not after it.
The full process this measurement sits inside, from the first call through to handover, is set out on how a build actually runs.
06
A worked example, labelled hypothetical
This is hypothetical. It is an imagined business, built only to show how the arithmetic works, not a company that exists or a client Praxis has worked with.
Picture a company that handles inbound service requests by phone and email. Before anything is built, its finance function counts a normal week and writes down the starting point: four people spend a combined 100 hours a week on these requests, most of it spent chasing status rather than resolving anything.
After the build, the identical measure is read again, from the identical records, by the identical finance function. This time the combined hours on the same work come to 40 hours a week.
The arithmetic: 100 hours minus 40 hours leaves 60 hours a week no longer spent on that work. That is the incremental change, in hours, measured against the starting point and nothing else. What those freed hours turn out to be worth in money earned or money saved is a question for the business's own finance function to answer in its own figures, not ours.
07
What happens if it produces nothing
If the after measurement shows no change against the starting point, using the same measure and the same source both times, the honest reading is that the work did not move, and we say so. We do not keep going on a result that is not there, and we do not pick a different measure after the fact to find one.
That does not mean nothing happened. Whatever was built stays with the business: the automation itself, the documentation, the training given to the people who use it, and the written record of what was tried and what the measurement showed. None of that is taken back because the number came in flat.
We stop when the agreed measurement shows nothing moved, not on a date we pick because the project is taking longer than we hoped. That is the point of fixing the measure before the build: the end of an engagement is read from the records, not argued from a position.
08
What you spend regardless
None of this comes without a cost to you. Your time is real, and so is what the automation asks of the business around it.
You spend your own time on the call and the diagnostic, and your finance people spend time agreeing the measure and signing off the starting point. We need real access: to the systems, the records, and the people doing the work, not a summary written up for us. Someone has to pay attention while the automation runs beside the old way, because a system nobody watches does not get trusted or fixed when it is wrong. And there is internal change involved almost every time, because a process that has run one way for years does not absorb a new step on its own. None of that is agreed on a call, and all of it is real.
Questions
What should you know before the first call?
- Why are the commercial terms not published on this page?
- Because they are agreed case by case, on the call, once the team has seen the work. A structure printed on a web page is chosen before anyone has looked at the business, and the work itself is the only honest basis for scoping an engagement to it.
- Who decides what counts as the starting measurement?
- The business's own finance people choose the measure, drawn from records the business already keeps, and it is fixed in writing before any building starts so neither side can change it once the after number is known.
- What happens if the business has never tracked the numbers a measure would need?
- Tracking starts from an agreed date before the build begins, and the starting point is whatever that new, simple record shows over an agreed period. A starting number is never invented from history that was never recorded.
- What if the automation gets built and it does not actually save time or make money?
- Then the same measure taken the same way afterwards shows no change, the team says so, and the engagement stops there rather than being argued into a result. Whatever was built, documented and trained still stays with the business.
- Who does the work once an engagement is scoped?
- The team does. Praxis is founder-led and draws on a team of 20 senior practitioners across the disciplines the firm advises on. The team sits inside the work, builds the automation, runs it beside the old way until it is trusted, and hands it over with written notes.
- What does the business actually have to give to make this work?
- Time from you and your finance people, real access to your systems and the people doing the work, and attention while the automation runs beside the old way. None of that is agreed on a call, and all of it is genuinely needed for the arrangement to work.
Where this leads
This page is one part of the whole offer: the AI Praxis puts into your business.
Ask what your starting point would look like
The fastest way to find out what this could be worth in your business is the free call. Bring what the work costs today, however roughly you know it, and we will tell you what a measure for it could look like and what we would build first.
To arrange the call, email praxisconsultingco@gmail.com with a time window that suits you. The reply confirms it; there is no booking page.
No obligation · a scoping conversation first