The agentic AI trust bottleneck is the reason most enterprise AI projects stall in pilot. Anudeep Parhar, Chief Operating Officer for Digital at Entrust, joins Alan Shimel to explain what that bottleneck really is. Furthermore, he lays out how enterprises can finally break through it.

About Anudeep Parhar

Anudeep is a three-decade technologist who has built, sold, bought and implemented enterprise software. Consequently, he brings both sides of the buyer and vendor equation into the conversation. Today, he leads Entrust's digital operations and its new AI initiative.

Inside the agentic AI trust bottleneck

Alan frames the harder question. Cost is the first bottleneck for enterprise AI, but the agentic AI trust bottleneck is a very close second. Between 70 and 90 percent of agents built today never reach production. In addition, roughly 20 percent of that shortfall traces directly to trust concerns.

In regulated industries the picture is worse. Financial services, healthcare and government workflows push the failure rate to 50 or 60 percent. Meanwhile, security teams cannot approve what they cannot audit. As a result, the agentic AI trust bottleneck holds real productivity gains hostage in the lab.

The trust plane enterprises are missing

Anudeep argues the missing layer is a trust plane. Enterprises must prove intent, consent and action for every agent and sub-agent. Therefore, security officers, risk teams and auditors can replay what an agent did and defend it after the fact. Explore more AI coverage and browse the latest TechStrong TV interviews.

Introducing the Entrust Trust Accelerator

Entrust just launched the Trust Accelerator to help customers define that trust layer together. Visit entrust.com/accelerator to join the program. Otherwise, talk to your Entrust representative for the on-ramp that matches your stage.