IDC VRT in partnership with Google Cloud & Salesforce

Virtual Roundtable

Scaling AI Past the Hype: Build the Foundation Agents Can Trust

October 29, 2026 | 3:00 pm ET / 12:00 pm PT

Logo Logo

AI agents are spreading faster than the foundations beneath them. According to IDC research, 95.3% of enterprises already run at least one agent workflow in production, yet only about one in three has a genuinely cross-functional multiagent solution, and fewer than half convert most proofs of concept into production.

 

The market is responding with architecture. Production agentic AI reaches 65% among organizations with high data product maturity, versus 14% at low maturity, and IDC predicts 45% of Global 500 enterprises will require a cloud-based headless application architecture by late 2027.

 

This IDC-moderated virtual roundtable looks at what it takes to scale past the hype: unifying disconnected data, applying one set of governance rules to every agent and tool, and deciding how much autonomy agents should have before they act on your core systems.

~80x

IDC forecasts active AI agents to grow from 31 million in 2025 to 2.45 billion in 2030, so the foundation beneath them must scale.

Retail and Services (431.5 million) and Financial Services (383.3 million) lead the 2030 Americas forecast of roughly 1.4 billion active agents, and every sector, including Public Sector at 95.3 million, deploys at scale, which makes shared foundations more important.

Each sector’s agents will touch very different systems. IDC finds organizations already average 22.3 data sources across 14 repositories, so unifying data and applying consistent governance across them is a prerequisite for scale, not an afterthought.

 

Source: IDC Worldwide Agentic AI Adoption Guide (IDC_P47511), July 2026; Americas, active agents in millions, 2030 forecast.

Discussion Themes

Four Conversations to Scale AI Agents Past the Hype

Where AI Scaling Stalls

Pilots are plentiful, but fewer than half of organizations convert most proofs of concept to production. Where architecture, legacy systems and fragmented data hold agents back, and how peers are sequencing the foundation before adding more agents.

 

 

One Set of Rules for Every Agent

Only about 35% of identity programs include AI agents today. How confident leaders are that every system, agent and tool touching customer data enforces the same governance, security and access rules, and who is accountable.

Freeing Business Logic and Data

Much business logic and data remains locked inside traditional applications. What it takes to open that foundation through reusable, open architecture so agents can work across any surface without rebuilding backend systems.

Trust Boundaries for Autonomous Action

68% of organizations are discussing more agent autonomy, and 29% already let agents decide in most areas. Where leaders draw the line on read and write access to core systems, and which actions still need a human in the loop.

What To Expect

A live video meeting led by a skilled host/moderator

An intimate gathering of senior leaders from diverse industries, grappling with similar issues

A secure, private conversation open only to a carefully vetted group (no recording, no playback)

Your hosts

Partners

Sponsored By