Every large organization has lived through this project at least once. The automation initiative delivers on time and on budget. The dashboard turns green, the steering committee applauds, the case study is written. And a year later, nothing has actually changed: the customers are no happier, the backlog is no shorter, the people are no less busy. The process runs faster — but it is still the wrong process.
Peter Drucker named this trap as early as 1963, in his Harvard Business Review essay “Managing for Business Effectiveness”:
“There is surely nothing quite so useless as doing with great efficiency what should not be done at all.”
— Peter Drucker
His line rests on a distinction that sounds trivial and is anything but: effectiveness is doing the right things; efficiency is doing things right. Efficiency asks how fast, how cheap, how automated. Effectiveness asks the more uncomfortable question — why does this work exist in the first place?
The systems theorist Russell Ackoff later sharpened Drucker’s point into a warning. Efficiency at the wrong thing is not merely useless, he argued — it makes you worse:
“The more efficient you are at doing the wrong thing, the wronger you become. It is much better to do the right thing wronger than the wrong thing righter. If you do the right thing wrong and correct it, you get better.”
— Russell Ackoff (quoted in The Wall Street Journal, November 2009)
In the age of AI, this warning has become urgent. AI is the most powerful efficiency engine we have ever attached to office work. It writes, summarizes, extracts, routes, and responds at a speed no human team can match. Which means: if the underlying process is broken, AI will now produce the wrong outcome faster, more consistently, and at scale. Efficiency amplifies whatever it touches — including the mistakes.
A large share of what our processes handle every day should not exist at all. John Seddon, who translated systems thinking into the Vanguard Method for service organizations, calls it failure demand: “demand caused by a failure to do something or do something right for the customer.” The second call because the first one wasn’t resolved. The complaint about the letter that wasn’t understood. The rework loop between two departments that were never designed to hand over cleanly. Automating failure demand does not create value. It industrializes waste.
To be clear: efficiency is not the enemy, and neither is automation. Slow, manual processing of routine work serves no one — and routine work is exactly what AI excels at. The point is not whether to automate — it is when.
The order decides everything.
First understand the system end to end: what do customers actually ask for, where does failure demand come from, and where does the flow break? Then fix it — remove the causes of failure demand, smooth the bottlenecks, redesign what needs redesigning. Only then automate, and with a much better question than “what can the machine take over?” What should we automate? What should we augment, so that human judgement stays in the loop? And what must remain genuinely human — the difficult conversation, the ambiguous case, the moment that needs empathy rather than throughput?
Understand, fix, automate — in that order. Reverse it, and you will get very efficient at doing the wrong thing. And as Ackoff warned: the better you get at it, the wronger you become.
Part 1 of the series “Empowering Transformation — Notes on Change That Sticks”. Next: why copying what success looks like is not transformation — it’s cargo cult.