# Would You Reject DRS Because It Got One Decision Wrong?

One bad AI interview call can put the whole idea on trial. Like DRS in cricket, the useful question is whether technology helps people make better hiring decisions than they would without it.

**Author:** Achin Agarwal · **Category:** Innovation · **Read time:** 3 min read

A close LBW decision goes to DRS. The replay runs, the ball-tracking appears, and half the room decides the technology is useless.

Curiously, this tends to be the half whose team lost the decision.

We can argue about that particular call for the rest of the evening. But there is a bigger question hiding behind it: would we rather go back to making every decision without the replay?

That is where the conversation gets interesting.

DRS does not have to make cricket error-free to be useful. Its value comes from the mistakes it helps us avoid, weighed against the ones it still makes. One bad call matters. It just cannot tell us the whole story.

I think we often forget this when we talk about AI.

An AI interview report gets one assessment wrong, and suddenly the entire idea is on trial.

“If it can make this mistake, how can we trust it?”

Fair question. But the human interviewer sitting next to it can make mistakes too. Two interviewers can hear the same answer and reach different conclusions. One sees confidence; the other sees bluffing. One likes a candidate’s directness; the other calls it arrogance.

Somehow, we find this easier to live with. We call it judgment.

We are familiar with human judgment. We have worked around its quirks for years. A machine’s mistake feels different because we expected the machine to remove uncertainty altogether.

That is quite a job description. Even the person who wrote it probably cannot meet it.

The self-driving car debate has a similar tension. An accident involving an autonomous car deserves serious scrutiny. But an accident alone cannot tell us whether autonomous driving is safer or more dangerous than human driving. That comparison depends on what happens across comparable journeys, including the accidents that never happen.

The same is true in hiring. We notice the candidate the AI got wrong. The capable person our old process quietly overlooked is much harder to spot.

None of this gives AI a free pass. A system that makes worse decisions is no improvement, however impressive its demo. And “humans make mistakes too” is a poor excuse for a bad product.

But there is a real cost to comparing a new tool with an imaginary version of the old process in which nobody ever got anything wrong.

For me, this is the interesting question behind Xquizit: can technology help people make better hiring decisions than they would make without it?

There is room in that question for evidence, disagreement and human judgment. There is even room for the technology to be wrong sometimes.

If perfection were the price of admission, the human interviewer would still be waiting outside.

**Want hiring decisions you can stand behind, even when a single call is messy?**

[Start Your Free Trial →](https://xquizit.in/pilot)

*14-day free trial · No credit card required*