Departments · Management and controlling AI for Management and Controlling: Reports, KPIs and Decision Papers

Reports and KPIs: six tasks at a glance

These are typical workflows, not client projects. In each one the AI prepares, and management or controlling checks and decides. Some of this is covered in more depth by our articles on meeting notes with AI and AI use cases for business.

1. Monthly and management report with commentary

The report is built from the same sources as last month, but the commentary has to be written again every time.

• Input: the reports from accounting, ERP and CRM as at the reporting date and last month's report.
• AI prepares: code calculates the KPIs and the differences from last month and from plan. The AI writes a commentary from them that names the main changes, and gives the source for every figure.
• Controlling checks: figures and commentary, and adds what is known only in the house.
• Output: a report draft with figures, charts and commentary.

2. Bring KPIs from several systems together

Revenue is in the ERP, open items in accounting, the pipeline in the CRM, and no figure matches another.

• Input: exports or interfaces of the systems.
• AI prepares: joins the data on common keys, calculates the KPIs with code and shows where systems differ and why (reporting date, period accrual, gross or net).
• Controlling checks: the differences, and decides which source applies.
• Output: a KPI overview with source and reporting date for each value.

3. Keep contract deadlines and obligations in view

Notice periods, renewals and payment obligations are scattered across contracts.

• Input: contracts and agreements from the filing system.
• AI prepares: reads term, notice period, renewal and payment obligations, enters them in a deadline list and reminds in good time.
• Management checks: the contract text before any decision. The AI shows the passage.
• Output: a deadline list with reminders and passages.

4. Minutes and decisions from meetings

Decisions and tasks get lost when the minutes are written days later.

• Input: the recording or notes of the meeting, if everyone involved agrees.
• AI prepares: writes minutes with decisions, tasks and owners and flags passages where it is unsure.
• Participants check: the minutes before they are sent.
• Output: minutes and a task list with deadlines.

5. Decision papers

Before an investment or a hire, figures, alternatives and risks should be on one page.

• Input: quotes, calculations, reports and your question.
• AI prepares: sets options side by side with the figures from the documents, names assumptions and open points and flags what it cannot support.
• Management checks: and decides. The paper is a summary, not a recommendation by the AI.
• Output: a decision paper with references to sources.

6. Cash forecast

When money comes in and goes out is spread across open items, contracts and tax dates.

• Input: open items, payment terms, contracts and tax dates from your systems.
• AI prepares: code calculates the expected cash in and out for each week from the data. The AI describes the trend and points to bottlenecks.
• Management checks: assumptions, such as customers' payment behaviour, and adjusts them.
• Output: a cash forecast with the assumptions behind it.

Programs and sources: connected to what you already use

Reports rely on several systems. The table lists the usual cases. Which route your system offers, we clarify before the quote.

The names are for orientation and do not mean a partnership. Where a system offers no interface, we work with export and import, a watched folder or email.

Connect, do not replace

You do not need a new reporting tool, and we do not sell one. The AI reads from your systems and writes the report that someone used to piece together.

• Your source systems stay in charge. Figures arise there and are only read by the AI.
• Your report format stays. The report looks as before and arrives earlier.
• One report at a time. We start with one report, for example the monthly report, and run it alongside the existing work first.
• If you want a dashboard, we build it on your data. It does not replace your source systems.

Approvals and traceability

A management team has to be able to rely on its figures. So before the start we agree how every figure comes about and who sees it.

• Source and reporting date: Every figure in the report names its source and its reporting date. If either is missing, the figure is not in the report.
• Calculating with code: KPIs are calculated by the program, not estimated by the AI. The AI describes what the figures show.
• Four-eyes principle: The draft goes to controlling or management before it is passed on.
• Permissions: You decide who sees which figures. The AI never shows a person more than they could see in the source systems.
• Differences: Where systems deliver different values, the AI names both and does not invent an average.
• Log: Every draft is recorded with sources, time and reviewer.

The table lists the main sources for orientation. Which duties apply to your company depends on legal form, country and articles of association.

This responsibility cannot be handed to a tool. An AI report is a working aid, and management remains responsible for the decision. For business figures sent to an AI service, the GDPR applies where people are affected, and so does your own confidentiality: a data processing agreement (art. 28) and processing in the EU. This overview is general information, not legal advice.

Data: KPIs and definitions

A report is only as good as the definitions behind the figures.

• Define KPIs unambiguously: What does revenue mean, gross or net, by invoice date or by date of service? We record the definitions in writing so that every report means the same figure.
• Reporting dates: Every value belongs to a reporting date. The AI compares only what applies at the same date.
• Data quality: Missing or duplicate postings in the source system are shown by the AI, but not corrected by it.
• Differences between systems: We explain them rather than smooth them over.
• Data protection: Reports contain figures on customers, suppliers and in part on employees. We process them on servers in the EU and agree retention and deletion with you. More under AI and data protection.

Your own programs for management

Some things are missing not in reporting but on the way to it. We build them as separate programs when standard products do not fit:

• Dashboard: the five to ten figures you really need, from your source systems, each with source and reporting date.
• KPI app for the phone: the state of revenue, open items and orders on the move.
• Approval app: payments and contracts for approval with the key details at a glance.

These programs complement your source systems, they do not replace them. More under app development.

Where to start

These three questions help with the assessment:

1. How long does the monthly report take from posting to management, and who writes it?
2. Which systems do the figures come from, and are there figures that never match?
3. Which report, deadline or set of minutes costs the most time?

If you send us the answers, we tell you which workflow is worth doing first and what has to stay with you. We begin with a trial run alongside the existing work.