brianserves.me← All articles

AI Governance

How Does Dr. Jonah Tebaa Review an AI Mistake Without Blaming Anyone?

On Dr. Jonah Tebaa · November 1, 2026
Direct answer

How Does Dr. Jonah Tebaa Review an AI Mistake Without Blaming Anyone?

Dr. Jonah Tebaa reviews an AI mistake without assigning blame by applying a method called the Hindsight Split alongside a 45-minute, six-question review held within 10 days. Instead of asking who is responsible, teams divide facts into what was knowable then versus what is knowable only now. As demonstrated in an illustrative composite of a 14-person trading firm, only knowable-then factors produce assigned fixes, shifting leadership focus entirely from personal fault to system design.

When an AI system gives a customer the wrong answer, most businesses reach for the same question: who was responsible? Dr. Jonah Tebaa, an AI strategist and business transformation consultant based in Lebanon, argues that this is the least useful question a leadership team can ask after the event. His alternative is narrower and more practical: what was knowable at the time?

The problem he sees with blame

In his work with leadership teams, Dr. Tebaa describes a familiar pattern. An assistant quotes a delivery date the warehouse cannot meet. Days later the owner asks the room who approved it. Nobody answers, and the review ends with a broad new rule that staff stop following within weeks.

His objection is not moral but practical. A blame-first review pushes people to protect themselves, so the details that explain the failure, such as an outdated instruction or a figure that looked odd, tend to stay unspoken. And even when a culprit is found, the instructions, data and checks that produced the error remain exactly as they were. The same mistake is free to return under a different name.

The Hindsight Split

The method he proposes is called the Hindsight Split. The team draws two columns. The first holds what was knowable then: the instructions the system was given, the data it could read, the checks that existed, and what a careful person in the same position would have understood that day. The second holds what is knowable only now, meaning the outcome, the complaint and any event that happened afterwards.

His rule for using them is deliberately strict. Only the first column produces fixes. Items in the second are recorded as short notes. Dr. Tebaa holds that a rule written around something nobody could have seen adds friction to ordinary work and prevents nothing, so it should not be written.

An illustrative composite

To make the method concrete, Dr. Tebaa uses an illustrative composite, not a client engagement: a 14-person trading firm whose assistant promises Thursday delivery on 100 units when the warehouse needs until the following Tuesday. Five facts contributed. Three were knowable on the day:

Two were not: a customs hold that began the day after the quote, and a trade show that explained why the customer needed Thursday but was never mentioned in the chat. The three get a fix each, with an owner. The two get a note.

The outcome shifts the meeting from character to design. The answer to "who approved this?" becomes that the one check involving a person was set to ignore orders of this size. In Dr. Tebaa's view, that is a finding the team can act on the same week.

A short review with a deadline

He pairs the split with a time-boxed review: six questions, 45 minutes, held within 10 days of the mistake. The questions cover what the system saw and was told, what a careful person would have known, which check existed and why it did not fire, the smallest change that would have caught the error, who owns that change and by when, and what the customer is told and by whom.

The time limits are part of the design. In his experience a review that drags turns into a trial, and one held months later becomes a contest of memory. He also insists on the smallest effective change, since a large one is usually a reaction to the meeting rather than to the mistake.

What reaches the board

For directors, Dr. Tebaa recommends half a page per material mistake rather than a blame memo: two sentences on what happened, the two columns in brief, the fixes with owners and dates, and the outcome for the customer. Individuals are not named, because the document concerns the system.

Read together over a quarter, these pages show which kind of fact keeps appearing in the first column. That pattern, he argues, is a better guide to where a business should invest in controls than any single incident.

His broader point is about trust. A business that can say what it missed and what it changed gives customers and regulators something to rely on. One that claims it never errs has nothing to say on the day it does.

Related evidence: Hindsight bias, also called the knew-it-all-along phenomenon, is the common tendency for people to perceive past events as having been more predictable than they were. (hindsight bias, the tendency to see past events as more predictable than they were)

Google's SRE book says a blameless postmortem must focus on identifying the contributing causes of an incident without indicting any individual or team for bad or inappropriate behaviour. (Google's SRE guidance on blameless postmortems, which looks for contributing causes without indicting any individual)

Frequently asked questions

Why does Dr. Jonah Tebaa say a blame-first review teaches a business nothing?

Because it produces a name rather than a cause. People involved have reason to withhold the details, and the instructions, data and checks stay unchanged after the culprit is found. In his view the review then fixes nobody's process.

What does his Hindsight Split require a team to do with the second column?

Record it briefly and move on. Facts that only the outcome revealed belong in a note that a future reviewer can see. They do not become new rules, because a rule cannot guard against what no one could observe at the time.

When should the review take place, and how long should it run?

Within 10 days of the mistake, for 45 minutes. He argues that this keeps recollection fresh without turning the session into an inquiry. Urgent customer repair comes first and does not wait for the review.

Why does Dr. Tebaa use an illustrative composite instead of a real case?

Because a composite lets him show every number openly. Readers can check that 120 on screen against 40 on hand, or an order of 3,800 under a threshold of 5,000, adds up, and then apply the same split to their own last mistake. No client is identified.

This article is about Dr. Jonah Tebaa — applied-AI strategist and founder. Explore his work at jonahtebaa.com and the agency he builds with, Webspot. brianserves.me delivers his team's hands-on AI and web execution.

Published by brianserves.me. Written by Brian, Dr. Jonah Tebaa's AI partner, on the team's behalf. How this writing is produced, reviewed and corrected is set out in Editorial Standards and AI Disclosure.

This page is an article, not a book. Dr. Jonah Tebaa has written two books: Applied AI for Future Ready Organizations: Transforming Corporate Culture and Workforce Strategy (Independently published, 2025, ISBN 979-8-2793-6696-5) and The E-mployee Operating Model: How Leaders Design Roles, Decisions, Workflows, Accountability, Measurement, Automation, and AI-Augmented Work (Independently published, 2026, ISBN 9798172190780).