American Express exec outlines 4 strategies for innovating in the agentic AI era

At a glance:

• Luke Gebb emphasizes curiosity, risk-taking, and partnerships for AI-driven innovation.
• Amex is developing tools like the Agentic Commerce Developer Kit and Agent Purchase Protection.
• Agentic commerce is expected to grow significantly by 2027, with early interactions starting this year.

Four strategies for AI-era innovation

Luke Gebb, head of global innovation at American Express, has spent 25 years navigating the financial services giant's evolution through digital transformation. Leading Amex Digital Labs, Gebb focuses on fostering a culture where innovators thrive by embracing curiosity and calculated risks. He argues that successful innovation requires a mindset shift away from incremental improvements toward bold experimentation.

"It's about getting a culture that people love, where they like to be together, and have the freedom to work on new things," Gebb told ZDNET. His approach emphasizes collaboration over isolation, ensuring that the lab integrates with business teams rather than operating as a separate R&D entity. This philosophy drives four core practices: continuous learning, deep technical understanding, willingness to fail, and strategic relationship-building.

Building the foundation for agentic commerce

Gebb's team is actively preparing for the rise of agentic commerce, where AI agents make purchasing decisions on behalf of consumers. Their efforts center on three pillars: enabling seamless payments, integrating with large language models (LLMs), and developing proprietary mobile and web experiences. The recently launched Amex Agentic Commerce Experiences Developer Kit aims to create trusted transaction frameworks, while Amex Agent Purchase Protection safeguards customers from AI-driven errors.

These initiatives reflect a broader industry trend toward AI-mediated commerce. As Gebb notes, card members will increasingly discover products through LLMs and delegate purchasing to agents. The company is also embedding booking capabilities into AI platforms, allowing users to seamlessly transition from recommendations to reservations within their preferred digital environments.

The road ahead for agentic commerce

While timelines for agentic adoption vary, Gebb anticipates early use cases emerging by late 2024, such as automated restocking or priority access to new products. By 2027, he expects agentic commerce to become mainstream, with average users engaging in routine AI-assisted transactions. However, the transition will likely blend human oversight with autonomous systems, requiring businesses to balance innovation with trust and reliability.

The financial services sector is particularly well-positioned to lead this shift, given its existing infrastructure for secure transactions and customer data management. American Express's proactive investments in agentic frameworks signal a strategic bet on AI's transformative potential, even as uncertainties around adoption rates persist. As Gebb emphasizes, preparation is key: "No one knows for certain if agentic commerce will happen faster or slower, but what we do know is it absolutely will happen."