An AI readiness assessment of your data model, before a price is set.

We read the consolidation, the ERPs and the spreadsheets between them, name every risk in the data and challenge the roadmap. The build price is fixed on what the survey found.

Read about AI advisory Start a conversation

An AI readiness assessment for finance is a survey of the data model a group's numbers run through, done before any price is agreed and before anything is built. Constancia, an EPM consultancy, sells it as a separate piece of work with its own written output.

Reads: The consolidation, every ERP and the spreadsheets between them.
Produces: A risk list with an owner against each item, and a roadmap challenged line by line.
Then: A fixed price for the build, set on what the survey found.

The assessment reads 6 things, and the documentation of them is never the same as what runs.

01
The consolidation as it runs
We sit through a close and read the rules HFM, BPC or the workbook applies, including the ones nobody still at the company has read.
02
A ledger from every acquisition
Each ERP exports in its own format and at its own grain, and we record what each one will release.
03
The spreadsheets between them
The mappings, the eliminations and the last mile of the close usually live in workbooks, and we list every one and who can open it.
04
1 word, 2 definitions
Gross margin, headcount and revenue by customer are read in every report that carries them, and each definition is written down beside its rivals.
05
Who owns each definition
We ask who would sign each definition if the auditor challenged it, and record the ones where the answer is nobody.
06
The close, day by day
The calendar is drawn as it is run, with every handoff from 1 person to another marked on it.

A price set before the survey is an estimate, and the estimate moves in month 4.

Scoping usually happens in a requirements workshop, because that is when the budget is set, and nobody has yet opened the data model.

When the third ERP exports at a different grain from the first 2, the finding arrives as a change order. The person who signed the business case then explains it to the board.

The assessment moves that discovery to the front, so the finding is in the fixed price before the client signs.

The assessment has a scope boundary, and both sides of it are written down.

In scope

Everything the group's numbers pass through before they reach the board.

The consolidation, the ERPs, the planning model, the spreadsheets that join them and the people who own each definition. Anything a model would later read is read first by us.

The data model: entities, accounts, periods and hierarchies
The close calendar and every handoff on it
The roadmap the board has already seen, challenged line by line

Out of scope

The tool choice, the ERP and any AI outside the finance data estate.

We do not choose the platform during the assessment, because the platform is a capability and the data model decides what it can carry. AI outside the finance data estate sits with another adviser.

ERP selection and implementation
A maturity score against a generic scale
A price for the build, until the survey is finished

A programme scoped before the model is read and 1 scoped after it differ in 5 places.

How an EPM programme scoped before the data model is read compares with 1 scoped after the assessment.
	Scoped before reading the model	Scoped after the assessment
When the price is set	At the budget round, from a requirements workshop.	After the survey, on what the survey found.
Where a surprise in the ledger lands	In a change order in month 4.	In the risk list, priced before the client signs.
Who reads the data model first	The build team, after the contract is signed.	The consultants who will build on it, before any price is agreed.
What the roadmap is based on	The software demonstration and the workshop's wish list.	The data model as it runs, in the order it can bear.
Who owns each definition	Decided during the build, by whoever is in the room.	Named in the survey, with the gaps listed.

Questions finance teams ask about an AI readiness assessment.

What is a readiness assessment?

A readiness assessment is a survey that establishes whether an organisation can support a change before money is committed to it. In finance the change is usually a model or an agent, and the thing surveyed is the data the model would read.

What are the 3 pillars of AI readiness?

Most frameworks name data, technology and people. In a finance function the data carries the other 2, because a model built on 2 definitions of gross margin produces 2 margins.

How to evaluate AI data readiness?

Ask 3 questions of every figure the board reads: which ledger it came from, who owns its definition and who can rerun it. The assessment asks them of the consolidation, the ERPs and the spreadsheets in turn.

Is your data ready for AI?

Finance data is ready when a number traces to its ledger, every definition has an owner and the close needs nobody's workbook. The assessment tests those 3 conditions entity by entity and lists where each one fails.

How to make data ready for AI?

Put the close, the consolidation, planning and reporting on 1 data model, agree the definitions and give each 1 a named owner. The build comes after that, and the assessment sets the order the model can bear.