INNOVAREModule 6 · AI Project Management

Case 2: The Number That Lied

A bank cut 45 jobs on the strength of one AI metric. The metric was wrong, and the workers knew it. How to measure an AI project so the number serves reality — not the decision.

July 2026 · Case 2 of 6
As you read — hold this question

When a measure becomes the target that justifies a decision, does it still measure anything real?

45 roles
people at Australia's biggest bank told an AI voice-bot had made their jobs redundant — on a number that was wrong.

In July 2025 the Commonwealth Bank told 45 call-centre staff their roles were redundant because a new AI voice-bot had cut incoming calls. The people on the phones knew the calls were not going down — they were going up. Within weeks some of the same workers, just told they were surplus, were asked back on overtime because the bot could not handle the hard calls. The Finance Sector Union's Julia Angrisano said the bank had been caught dressing up job cuts as innovation. The bank admitted in writing that its assessment 'did not adequately consider all relevant business considerations' and the roles were not redundant.

The mechanism
Quiz: Objectives & KPIs

Goodhart's Law, at a bank

There is a name for how a number goes wrong like this: Goodhart's Law — when a measure becomes a target, it stops being a good measure. The moment "calls down 2,000 a week" became the justification for the cuts, it stopped measuring reality and started serving the decision. The technology worked. The failure was a management decision made on the wrong number.

Measuring an AI project properly

SMART goals, OKRs, and the KPIs that actually matter

A good goal is slippery, so the profession built a shorthand: SMART — Specific, Measurable, Achievable, Realistic, Time-bound (George Doran, 1981). An AI goal without a deadline cannot be measured. To measure, you reach for OKRs (qualitative Objectives and their measurable Key Results) and KPIs (the quantitative gauges). Each KPI ladders up to a macro objective — e.g. Objective: increase efficiency through automation; KPIs: reduce manual processing time by 50%, or reach 90% accuracy on AI-generated reports.

Here is where AI projects go wrong: the obvious metric — model accuracy — is not the one that matters. Your course lists the KPI categories that do:

KPI categoryWhat it measures
Return on investmentFinancial return from added revenue + cost savings vs. AI investment
User adoption & engagementUser adoption rate; feature utilisation — are people actually using it?
Personalisation effectivenessPersonalisation conversion rate; personalisation accuracy
Brand perception & loyaltyBrand sentiment (before/after); brand loyalty
Stakeholder & customer satisfactionStakeholder engagement; customer feedback
What is missing from that list
Not one category is "the model scored 95%." A model can be accurate and still be failing the business. Measurement chosen carelessly is how good teams talk themselves into terrible decisions — exactly what happened at the bank.
The wider pattern

Why most AI projects fail (Bernard Marr)

Up to 80% of AI projects fail to deliver their intended value. Marr (2025) names three pitfalls, and none of them is the technology:

1
Technology first
Rushing to implement AI because competitors are — instead of starting from a clear business problem.
2
Data quality
"Garbage in, garbage out." Organisations vastly underestimate the work to prepare and maintain clean, relevant data.
3
People & change
Teams are not prepared for new ways of working. The best technology fails if people don't understand it or resist it.

What successful organisations do differently: start with a clear business case and measurable objectives; invest in data quality first; do change management and training from day one; and start small with a pilot before scaling.

Take this away

Measure the value, not the maths. A metric that becomes the justification for a decision stops describing reality. Define SMART objectives, ladder KPIs up to real business outcomes, and never let model accuracy stand in for whether the project actually helped anyone.

Quick recall — without looking back

Test yourself on this case

Question 1 of 3

What does SMART stand for, who introduced it and when, and why must an AI goal be time-bound?

SMART = Specific, Measurable, Achievable, Realistic, Time-bound. George Doran introduced it in 1981. An AI goal must be time-bound because a deadline is what allows performance and objective-achievement to actually be measured; without one the goal is only a wish.
Question 2 of 3

Distinguish OKRs from KPIs, with an example.

OKRs are qualitative Objectives with measurable Key Results describing what to achieve; KPIs are the quantitative metrics that signal success. Each KPI ladders up to a macro objective. Example — Objective: increase efficiency through automation (qualitative); KPIs: reduce manual processing time by 50%, or achieve 90% accuracy on AI-generated reports (quantitative).
Question 3 of 3

Name the five KPI categories the course lists for AI implementation, and explain why model accuracy is not among them.

ROI; user adoption & engagement; personalisation effectiveness; brand perception & loyalty; and stakeholder & customer satisfaction. Model accuracy is absent because a model can score highly and still fail the business — the KPIs measure delivered value (money, usage, experience, trust), not the maths. Measuring the wrong thing (Goodhart's Law) is how the CommBank cuts were justified on a number that did not reflect reality.

Module 6 Videos

Module 6 · Short · The Report That Made Itself Up
Module 6 · Long Form · Why AI Projects Fail

Sources

CommBank / FSU
Reporting on the Commonwealth Bank AI voice-bot redundancies and their reversal; Finance Sector Union (Julia Angrisano), 2025.
Module content
BUSN9049 Module 6 — SMART (Doran 1981), OKRs (Zokri n.d.), KPI examples. Flinders University, 2026.
Go Further
Marr, B. (2025). Why Most AI Projects Fail. YouTube (2:48).