The human judgment required by Gojek’s ex-CTO to scale AI tools

This article summarizes an episode of Tech Lead Journal’s video series featuring Ajey Gore, former group CTO of Gojek.

Ajey Gore, former group CTO of Gojek / Photo credit: Ajey Gore

AI can make teams work faster, but it also makes their mistakes harder to catch and fix. Ajey Gore, former group CTO of Gojek and former operating partner at Peak XV Partners, points out that AI is making standard technical skills much cheaper.

The real challenge for companies now is whether they can clearly define what “good work” looks like before AI produces too much of it. Because of this, using AI is just as much a leadership test as it is a technology test.

Human judgment is the new most valuable skill

As AI takes over repetitive tasks, the real value within a company shifts to people who can set high standards, spot risks, and decide if a generated answer is trustworthy.

Gore believes that verifying this output must become a core business skill, not just a quick final step.

“When expertise is becoming almost free, what becomes expensive? My answer is judgment becomes expensive,” he says. “We need to fundamentally start focusing on the judgment layer, because if agents can do good things fast, they’ll do bad things fast as well.”

Just creating a large amount of work is no longer impressive, because AI can quickly multiply bad ideas before anyone notices the error. Companies must reward employees who deeply understand what success looks like, rather than those who simply produce the most artifacts.

The takeaway is that deep human experience is becoming valuable again. The biggest payoff now comes from knowing exactly what problems to solve, a skill much harder to measure, and replace, than simple task completion.

Workers succeed by defining what matters

This shift also changes what it means to be a senior employee. Developers who insist on writing every line of code by hand are protecting a skill whose financial value is steadily declining.

Instead, Gore encourages engineers to let AI handle routine execution so they can spend more time understanding the problem, designing stronger solutions, and verifying the results.

“What you are getting paid for, and what you pay for, is always a skill. You will never pay for typing speed. Typing speed and some structured skills are gone. Now you have to define what is correct in domains as well.”

For businesses, the risk is paying for the wrong things. Teams might keep praising people for looking busy, even though the real value lies in planning, evaluation, and decision-making.

The best engineers will stop acting like simple task-finishers and become translators who turn vague ideas into clear plans that AI can build and humans can verify.

AI safety requires strict, traditional engineering habits

To save money with AI, redesign workflows instead of just firing people