Solving Hard Problems With Data And Software

In this DLD22 session, Ron Agam (SAP Signavio) and Noam Perski (Palantir) discuss the power of data analytics and how it could help to create an autonomous, intelligent enterprise.

The vision of autonomy in an organization is about “something that’s really more than just a specific function that can be automated”, Agam explains.

Today, the state of the art is “data, connectivity, plumbing, labeling”, the SAP manager says – but this typically requires a lot of human intervention still. “All the systems have data”, Agam says. “It’s just [that] you can’t really access everything at scale immediately.”

A crucial question, Perski notes is, what decisions in industry and government processes should be automated so that they do not need human intervention anymore. “Is it achievable? Is it desirable? And how far away from that are we?”

The coronavirus crisus provided valuable insights in this respect, Perski says, adding that Palantir was “very involved” in assisting the British National Health Service (NHS) respond to the pandemic.

Effectively managing hospitals and creating a vaccination campaign required bringing various data sets together, Perski explains. “So you started to see the emergence of a reasonably autonomous system where the machine is actually able to understand how we think this phenomenon is going to develop.”

Transparency is crucial as algorithms become decision makers, Agam and Perski agree.

“Ultimately it becomes a question of who’s designing the system?”, Perski says. “And how much do they believe in certain values or certain people?”

Perhaps the ultimate goal should not be entirely autonomous systems, Agam suggests, but rather an “augmented information worker” who can “work in a better, more productive environment” thanks to the support of intelligent software.