Profitability Trust Score™

An independent score for the model your board is about to act on.

Your team, or an AI tool, has built a cost and profitability model. Before anyone signs off on a pricing change or a divestment, we test the model across seven dimensions and return a single score, 0 to 100, with a certified report. CFOs use it to defend the number. Boards use it to approve. PE buyers use it to price the risk.

Request a Trust Score See the report

From 2 weeks · Fixed fee · Signed by a senior partner

Sample model

01What it is

The audit of an AI cost model. Independent, by construction.

A Trust Score is a structured test of seven dimensions of a profitability model. We look at the data going in, the logic that allocates costs, how the model behaves when reality moves, and whether the answer is connected to anything that has happened to the business. Each dimension is scored. The whole is scored. The result is a number, a report, and a list of fixes ranked by impact.

We do not sell the software the score validates. The score is the only thing on the table.

02The scale

0 to 100. Four bands. One confidence threshold.

A global score and a score per dimension. The 75 line is the threshold we tell CFOs to act on. Below it, the model still informs the conversation. Above it, you can decide.

00 to 49

Unsafe

Material issues across multiple dimensions. The numbers will not survive contact with the data. Do not act.

50 to 74

Indicative

Direction is right. Magnitudes are not yet reliable. Useful as a working hypothesis, not as a board number.

75 to 89

Defensible

Material decisions can be made on the model. Findings can be defended internally and to a sophisticated buyer.

90 to 100

Audit grade

Robust to drift, traceable end to end, ready for a regulated process or PE due diligence.

03The seven dimensions

Seven angles on the same question: can I act on this number?

● Sample model · Global score 87 / 100

0192 / 100

Data quality

Completeness, consistency and recency of the data feeding the model.

0288 / 100

Traceability

Every euro on the output can be walked back to a transaction and a rule.

0385 / 100

Allocation logic

Drivers are economically right, not just arithmetically convenient.

0481 / 100

Drift

Does the model still hold when the period, mix or volume moves.

0586 / 100

Bias

Systematic over or under attribution by segment, channel or geography.

0684 / 100

Robustness

Survives a sensitivity test, a missing month and a noisy upstream system.

0791 / 100

Link to outcomes

Predictions reconcile to what actually happened, period after period. The honest test.

04How we test each dimension

The method, in plain language.

Enough to satisfy a board, not so much that we are publishing the test set.

01 Data quality

Profile against expected types, ranges and cardinalities. Tie out to the general ledger to the cent.

02 Traceability

Pick fifty output cells at random. Walk each back to the source row and the rule. Lineage report.

03 Allocation logic

Re-derive drivers from first principles. Compare to the model's choice. Flag economically odd picks.

04 Drift

Re-run the model on the previous two periods. Bound any walk in scores. Watch for cliffs.

05 Bias

Slice the residuals by segment, channel, region. Test for systematic over and under attribution.

06 Robustness

Pull one month, perturb the inputs by 5%, blank a driver. Measure how the result moves.

07 Link to outcomes

Reconcile model gross margin to actual gross margin, period by period. The honest mirror.

05The report

The artefact you receive: a certified report.

PDF and CostCtrl workspace. Three sections do the work: the cover with the score, the dimension breakdown, and a ranked list of fixes with expected impact.

Profitability Trust Score™

CERTIFIED

Profitability Trust Score, certificate of independent validation for an AI-built cost model.

CLIENTSample · manufacturer

MODELTDABC v3.2 · 1,247 SKUs

PERIODFY2025

Miguel Guimarães

Senior Partner · Lead Cost Engineer

Cover

01 Cover & certification

Score

Defensible

GLOBAL · 0 to 100

Data quality92

Traceability88

Allocation85

Drift81

Bias86

Robustness84

Outcomes91

Score

02 Score by dimension

Actions

Fixes, ranked by impact on the global score.

01 · HIGH+5.2

Add setup-time driver to short-run SKUs

02 · HIGH+3.8

Reconcile depreciation to fixed-asset register

03 · MEDIUM+2.4

Split logistics pool by channel

04 · MEDIUM+1.9

Add drift monitor to monthly close

Projected score after fixes: +13.9

Actions

03 Ranked actions & impact

06Who asks for it, and why

Three moments where the score does the work.

CFO

Before a pricing decision.

The team has a recommendation. The CFO can act on it only if the model behind it survives scrutiny. The score either ends the debate or sends them back to fix the two things that need fixing.

Board

Before approving a portfolio move.

Closing a line, exiting a customer segment, repricing the whole catalogue. A signed independent score makes the decision defensible in the minutes, and a year later when the auditor asks.

PE buyer

During diligence.

Sellers now ship AI built profitability decks. Buyers need to know what those numbers are worth. A Trust Score either confirms the valuation or hands a price adjustment lever to the IC.

07Scoping

What an engagement looks like.

Duration

2 to 3 weeks

Two for a focused model, three for portfolio scope.

Your time

2 to 4 hours

Per week, from a controller who knows the data.

Data we need

Financial & operational data

Read access, under NDA. ERP, Excel or BI dataset. We do the wrangling.

First step

30 min, free

A senior partner. No deck, no follow-up sequence.

08Questions a CFO asks first

The honest answers, before the call.

What is the Profitability Trust Score?

A certified, independent assessment across seven dimensions of how trustworthy an AI-built cost or profitability model is. Created by Cost and Profitability Consulting.

How long does an assessment take?

Typically two to three weeks, depending on the complexity of the model and the availability of data.

Do you validate models you did not build?

Yes. The Trust Score is independent: we assess any model built with AI, ours or a third party.

Does the Trust Score help with EU AI Act compliance?

Yes. The traceability, data governance and human oversight dimensions map directly to the EU AI Act requirements, and the report serves as auditable evidence.

How do I know if an AI-built cost model is correct?

Looking right is not enough. We reconcile the model with the financials, test the allocation logic and capacity, and check every number before it is used. The result is a score from 0 to 100 and a report that backs the decision.

Who issues the Profitability Trust Score?

Cost and Profitability Consulting, an independent firm that does not sell modelling software. That independence is what makes the score credible to boards, auditors and buyers.

●Profitability Lab

This thinking comes out of our Profitability Lab.

The Lab is the research unit where we build the framework and the field notes on AI, cost and profitability.

Visit the Profitability Lab

Start with the score

Bring the AI model. We will tell you, in three weeks, what it is worth.

A single fixed fee. A senior partner on every working session. A signed report at the end.

Request a Trust Score Back to AI Profitability

Ask us anything

usually replies in minutes

Hi. I can answer the quick questions about cost, method and timing right here. For anything specific to your business, I'll connect you with a CostCtrl specialist on WhatsApp.

Talk to a specialist

Free. No bot loops. Straight to a specialist.