A dashboard can train the system to value collapse

An operating dashboard can get smarter and still teach the business the wrong desire.

Hadto should take one practical lesson from The Last Economy. At macro scale, Emad Mostaque argues that GDP can count disaster recovery, sickness treatment, and attention capture as growth. The business-level version is easier to miss. A service company can reward more tickets, more touches, more recovered failures, and more owner intervention while calling the whole pattern improvement.

The problem is not measurement by itself. The problem starts when the system is allowed to optimize one target without naming the human values around it.

A busy loop can still be a collapse loop

Take a home-services dashboard. It may track booked jobs, technician utilization, callback resolution, financing conversions, and response speed. Useful numbers. But if the operating target becomes “keep revenue and throughput rising,” the system can quietly learn bad substitutes for health.

It can normalize rushed diagnostics because faster closes help the board. It can tolerate more callbacks because recovery work still fills the schedule. It can over-message customers because attention looks like engagement. It can keep the owner trapped in exception review because intervention saves the metric in the short run.

The dashboard stays active. The business gets weaker.

This is not abstract AI risk. It is a normal operating failure. The machine learns to harvest the damage it should have prevented.

Owners need protected values, not just targets

Hadto’s thesis is not that AI should produce more output. It is that AI should help create more capable owners.

That requires a harder operating contract than a target metric alone. Every serious dashboard should declare three things:

• the optimization target
• the protected human values
• the anti-goals

In a service business, the target might be faster and more reliable job flow. The protected values might be diagnostic honesty, customer trust, operator judgment, and a business another manager can run without rescuing every exception. The anti-goals should be stated just as plainly: do not increase callback demand and call it utilization. Do not increase owner interruptions and call it responsiveness. Do not stretch customer attention and call it engagement. Do not turn team exhaustion into apparent throughput.

Without those boundaries, the system will often choose the local win that damages owner capability.

The anti-goals are what make the metric safe

Many teams treat anti-goals as ethics language for later. They belong in the instrument panel from the start.

When a dispatch AI is rewarded for same-day bookings, the dashboard should also show whether first-time fix rates fell, whether callbacks rose, whether exceptions came back to the owner, and whether customer-contact volume increased without better outcomes. A collections workflow rewarded for recovered balances should also show whether protected member obligations were overridden, whether appeals rose, and whether the office is spending more human attention cleaning up machine aggression.

Those checks do not slow down the business. They stop the business from training itself on collapse economics.

The point is not to remove automation. The point is to keep automation from mistaking breakdown, dependency, and captured attention for value creation.

The right question for Hadto

When a dashboard improves, Hadto should ask one more question before celebrating: did the gain make the next owner more capable, or did it just make failure more harvestable?

That operator test is worth keeping in the AI economy.

A good system gives the owner more legible judgment, fewer rescue obligations, and a company another operator can inherit. A bad one gets excellent at scoring the consequences of drift.

AI should not just make the business busier. It should make ownership more transferable.

That is why every important metric needs protected human values and explicit anti-goals attached.

---

Source evidence used in this note: reviewed Emad Mostaque, The Last Economy, especially Chapter 4 (“The Dashboard for Insanity”) and Chapter 15 (“The Alignment Economy”), along with the completed local reading note and study ledger for this book and recent Hadto posts checked on 2026-05-08 to avoid duplicating earlier notes on green dashboards, reviewable decisions, and playbook change.