Insight · without a survey

Checking Load Assessment

What your systems produce, and who checks it. Two numbers that have never been side by side in your company, and a list of functions nobody decided on.

What it's about

Half past seven in the evening, forty-eight reports, one hour. The head of nursing hasn't done anything wrong. Still, she doesn't read the one half-sentence that would have mattered.

Over the past eighteen months something has happened in most organizations that nobody decided. A software update added functions that pre-sort, score or rank. It was in the release notes, and hardly anyone reads release notes.

At the same time, more arrives at the review points, the places in a process where someone checks and approves. The hours there have stayed the same. Checking that doesn't get done doesn't disappear. It lands on a person who neither ordered it nor noticed it, usually the one who signs at the end.

The Checking Load Assessment turns this into a number. It works only with data that is produced anyway, and it doesn't survey anyone. That's deliberate. In a survey, hardly anyone admits they can no longer manage a check.

You can book the Checking Load Assessment on its own or within the Rule Assessment. That also counts your review points and covers the decisions that already tie down your future.

Who it's for

• Managing directors and owners of companies with 40 to 600 employees who want to know where their systems pre-sort or score.
• Companies with a Management Baseline that want to check the self-assessment against system data.
• Supervisory and advisory boards and compliance teams who need to know what approvals really rest on before they formally approve management's actions.
• Works councils assessing co-determination under Section 87(1) no. 6 of the German Works Constitution Act (BetrVG), who need to know which systems score anything at all.

Typical occasions: before an AI policy, after a near miss in an approval, before investing in more tools, when preparing a works agreement.

How it works

1. Clarity Call and scope, 45 minutes. We decide which decision routes to measure. Two to four are enough, typically purchasing, incoming invoices, hiring and service tickets.
2. Data handover, half a day on your side. Exports from systems you already run: approval logs, ticket system, purchasing, change histories of your business systems. The analysis covers cases and review points, and no individual people. Where there's a works council, we agree the scope with it first.
3. Analysis without your involvement. We measure throughput at each review point, today and eighteen months ago. We also measure the change rate, the share of cases in which the review altered anything. The time per case assumes six focused hours a day. That figure shows how much time there is on paper, and whether it's enough depends on how hard the case is. In parallel, eighteen months of release notes from your business systems are read for one question: which function that pre-sorts, scores or ranks arrived unnoticed?
4. My reading time, stated in the report. The machine proposes candidates. I decide which of them are right, and the report states the hours that took and what I discarded.
5. Review call, 90 minutes. We set the numbers against your responsibilities: where is a decision owned by a person who, at this pace, has no time for it?

What you have in hand at the end

• The number pair for each review point: what arrives, against the time set aside for it, compared with eighteen months ago.
• The change rate. A change rate close to zero is a reason to look closer and doesn't settle anything yet. It can mean good work before the review or too little time for the review. The report names the review points where your people should take a sample to find out.
• The list of embedded functions with the date they appeared and the decision they affect.
• The conflict lines, where ownership and pace don't fit.
• A report you can put before a supervisory body, with the data basis and my checking time stated.

The numbers show you where you can leave more to a program, because the checking is good enough, and where you have to make time first.

Prices

All prices plus VAT where it applies. An existing report from the Management Baseline is credited against the scoping.

How it differs

• Management Baseline: it measures by self-assessment how your leadership team sees the organization. This assessment analyzes what your systems log. Read side by side, the two show the gap between the situation as reported and as built.
• IT audit or software inventory: it lists which systems are running and how they're licensed. This assessment is about the decisions they affect.
• AI readiness assessment: it asks how far a company has got with AI. The answer reliably comes from the project list and misses most of what is actually running.

What this offer can't do

The assessment shows who ends up with the checking that doesn't get done. Whether your company wants it that way is a value decision, and that decision stays with your leadership team.

And it depends on what gets logged. Where a decision route leaves no trace, the measurement finds nothing. That gap is a finding in itself, just rarely the one you ordered. Phone calls and corrections before approval aren't in the data either, so in the review call we read the numbers against what you know from daily work.

Where the method comes from

I describe the calculation behind it in Cheap to Make, Costly to Check (2026), the first volume of the series “Organization under AI.” Ten questions for introducing AI, each with a measure, are in the free white paper “What AI Tests in Your Rules”.

When the assessment shows more

Often the assessment shows more than too little time for checking. Behind an approval nobody can manage any more, there are usually unsettled decision rights, roles without a clear mandate or departments steering past each other. For larger companies the assessment can therefore be the start of a broader redesign of organization and management.

In the Rule Redesign your leadership team decides on review points and decision rights. In Organizational Design workshops set out how your management system will work. An organizational constitution puts the rules in writing, including limits and priority in a conflict. The annual mandate covers this whole path over one year.

Next step

The way in is a Clarity Call, where we work out which decision routes are worth measuring and whether the data is enough. To check alone first whether this fits your situation, try the situation analysis from “Cheap to Make, Costly to Check”. It takes fifteen minutes and is free on the series website.

How much time is really left for checking an AI suggestion can also be observed on a small scale before the assessment begins. The free app Clarity Loop records, for each of your own decisions, how long you checked and what you changed, and stores everything only on your own phone.