The AI-to-P&L Note · No. 5

Three AI questions for your next exco

Three questions tell an executive team whether an AI proposal is ready for a decision. One is about value, one is about the work, and one is about accountability. In South Africa, the third has a statute behind it.

Watch the 32-second version

1. Which P&L line will this move, by how much, and by when?

This is the value question. A good answer names one line, a baseline from your own data, a target and a date. A weak answer talks about productivity, innovation or “learning”. Those may be real, but they cannot be measured against the income statement, so they cannot be funded against it.

The question matters because the gap between experimenting and earning is wide. Boston Consulting Group’s 2025 research, The Widening AI Value Gap, found that 60% of companies report little or no value from AI, and only 5% achieve value at scale.

2. Which workflow changes, and who owns the change?

This is the work question. Value comes from work that stops, decisions that are made faster or better, and capacity that is redeployed. If the old process runs in full alongside the new tool, the cost stays and the benefit leaks.

A good answer names the steps that change and one executive who owns the redesign, not only the technology. If IT owns the platform, operations owns the process and finance owns the target, ask which of them owns the outcome.

3. Where does a human decide, and can we evidence it?

This is the accountability question, and in South Africa it is also a legal one. Section 71 of the Protection of Personal Information Act (Act 4 of 2013) restricts decisions based solely on automated processing of personal information, where the decision has legal consequences for a person or affects them to a substantial degree. The Act sets out exceptions and the safeguards that must then apply.

Credit decisions, claims and applications are obvious candidates. A good answer shows where a person decides, what that person can see, and the record that proves it, in a form you could put in front of the Information Regulator. A human reviewer who approves by default may not be the safeguard management believes it is.

If the proposal cannot answer all three, it is not ready for a funding decision. It may still be a good idea.
How to use the three questions

Send them to the sponsor a week before the exco paper is due, and ask for one paragraph on each. Fund the proposals that answer all three. Send the rest back with the gap named, rather than declining the idea.

The point

Exco time is the scarcest input in an AI programme. Three questions, asked the same way every time, turn it into decisions: fund, fix or stop. For a deeper read on the third question, see Note No. 2 on POPIA section 71.

Take the free AI Governance Check

Sources

  1. Boston Consulting Group, The Widening AI Value Gap, 2025: 60% of companies report little or no value from AI; 5% achieve value at scale.
  2. Protection of Personal Information Act 4 of 2013 (South Africa), section 71: Automated decision making. Summary above is a paraphrase; refer to the Act for the authoritative text.
  3. Information Regulator (South Africa): inforegulator.org.za.
Next step

Apply this to your organisation.

The Operating Model Diagnostic gives your executive team a prioritised, governed view of where AI can move the P&L. Pricing on request.

Request an Operating Model Diagnostic