Enterprises Face Challenges in Scaling AI Agent Adoption

Governance and Security Remain Top Concerns

CIOs across major enterprises, including PepsiCo and Principal Financial Group, are grappling with the challenges of implementing AI agents at scale. While pilot programs have surged from 37% to 65% in recent months (per KPMG's survey), widespread deployment remains at just 11%. Key hurdles include:

• Governance gaps: Traditional frameworks struggle with AI's nondeterministic nature
• Security risks: Gartner predicts agents could accelerate cyber threats
• Technical immaturity: Vendors lack interoperability and scalable controls

Enterprise Approaches Vary

PepsiCo's Athina Kanioura emphasizes developing internal policies amid absent government standards, while Lowe's Chandhu Nair cautions:

"We are taking a very cautious approach knowing the technology is early"

Other examples:

• Expedia Group: Partnered with OpenAI but remains skeptical of agent hype
• Ulta Beauty: Focused on foundational work before widescale implementation

Change Management Critical for Workforce Adoption

Employee concerns persist about:

1. Lack of human oversight
2. Accountability for errors
3. Potential bias

Principal Financial Group's Kathy Kay uses training to build trust:

"We talk about our enterprise risk AI framework so employees understand our governance"

Looking Ahead

• Gartner predicts 40% of agent projects will fail by 2027
• Early adopters report productivity gains (PwC found 26% budget increases planned)
• EY's Joe Depa stresses: "We need to certify agents are acting responsibly"

As PepsiCo's Kanioura notes, collaboration is key: "Sharing best practices pushes the tech community to develop real solutions."