Maximum Autonomy ≠ Maximum Value

The AI industry is racing towards autonomy. I’m not convinced that maximum autonomy is the right goal.

Take customer service as an example. AI can analyse interactions, recommend what to do next, support employees during a conversation, take over parts of a process or handle the entire interaction autonomously. But not every use case needs to move all the way to full automation.

If I order a taxi here in Hanover, I now speak to an AI. And I actually prefer it. It’s a simple request, I get what I need immediately and I don’t have to wait for someone to pick up the phone. There is very little value in having a human involved.

But that doesn’t apply to every interaction. There are situations where I want a human to handle the conversation, while AI supports in the background. In others, AI can take over individual steps such as authentication. And there are simple, repetitive processes where full automation is clearly the better experience.

I don’t think this is a question of human or AI. It’s more like a set of sliders that every company has to adjust for itself: Where does AI support? What can it take over? And where does a human create real value?

Customer service is just one example. The same question applies across an organisation. The answer will depend on the process, the risk involved, regulation, customer expectations and ultimately the company’s own DNA.

For me, the goal shouldn’t be maximum autonomy. It should be finding the right level of autonomy for each use case.

More autonomous doesn’t automatically mean better.

Where would you move the sliders in your organisation?

LINK