AI News/AI News

SAP and Infosys execs: treat AI agents as a platform, not pilots

In a sponsored MIT Technology Review podcast, SAP and Infosys leaders argue that data, governance and use-case discipline decide whether AI agents scale.

Key points

  • SAP's Thorsten Leiduck and Infosys' Satinder Pal Singh discussed enterprise AI agents in a sponsored MIT Technology Review Business Lab episode produced with Infosys.
  • Infosys and HFS Research surveyed over 500 Global 2000 enterprises and found only 14% have successfully scaled AI; 86% remain in early or pilot stages.
  • Singh recommends scoring agent use cases on impact, feasibility and scalability; Leiduck advises starting with a small set of high-value cases with defined success metrics.
  • Leiduck says SAP launched an autonomous enterprise framework to help companies move from pilots to scaled outcomes.

Why it matters for marketers: Marketing teams piloting AI agents get a checklist for picking use cases, setting metrics up front and fixing data and governance before scaling beyond isolated experiments.

In a sponsored episode of MIT Technology Review's Business Lab, produced with Infosys, SAP's Thorsten Leiduck and Infosys' Satinder Pal Singh described why many enterprise AI efforts stall. Their central argument: companies run isolated pilots on fragmented data, then find no clear path to scale. Both urge treating AI as an enterprise platform and operating model, not a set of side projects.

The speakers cited several figures. Singh pointed to an Infosys and HFS Research survey of over 500 Global 2000 enterprises in which only 14% had successfully scaled AI. Leiduck cited a BCG report saying 74% of organizations are stuck in pilot mode. These are the speakers' own citations in a sponsored discussion, so read them with that context.

For marketers and content teams, the practical advice is easy to borrow. Singh suggests judging each agent use case by impact, feasibility and scalability, and says a pilot that can't scale is just an experiment. Leiduck recommends starting with a small set of high-value cases and defining success metrics first, such as cycle time, accuracy or cost per transaction. He also notes that not every task justifies an agent, since human effort is sometimes more economical.

Both stress data and governance. Leiduck says AI depends on a connected, governed data foundation, and Singh warns that skipping governance creates a risk backlog that can force a disruptive stop and audit.

Next steps for teams: audit the content and customer data your agents would rely on, set access and monitoring policies before expanding, and pick one or two workflows with measurable targets. Treat vendor survey numbers as directional, since the speakers represent companies selling AI platforms and services.

Sources

This brief is our own summary. Read the full reporting at the source.

You read about AI automation. We run it.

DocFluence, built by Z Media, writes, designs, films and publishes for your brand every day, in your voice, from one login.