3349: IBM Consulting on Agentic AI and the Reinvention of Business Processes

Jul 17, 2025

Francesco Brenna, Global Leader of AI Integration Services at IBM Consulting, delves into the transformative power of agentic AI. He emphasizes the shift from passive assistants to intelligent agents capable of executing tasks, which requires rethinking business processes from the ground up. Data readiness remains a crucial challenge, and he introduces 'data products' to anchor AI behavior. The discussion also highlights the role of human-centered design and upskilling in integrating AI effectively while proposing standards like Model Control Protocol for seamless operations.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Agentic AI Transforms Workflows

Agentic AI shifts the focus from mere assistance to actual execution of work.
This evolution enables entirely redesigned workflows instead of just patching old inefficient processes.

ADVICE

Prioritize Data Readiness

Prioritize data readiness by modernizing data foundations with governance to support reliable AI agents.
Design data products that provide context and memory to ground agent behavior effectively.

INSIGHT

New Architecture for Agentic AI

Agentic AI requires a new application architecture of user experience, orchestration, and data layers.
This enables dynamic, conversational interfaces and seamless orchestration across multiple platforms grounded in reliable data.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

When AI enters the boardroom, it tends to arrive with big promises: productivity, automation, efficiency. But for Francesco Brenna, Global Leader of AI Integration Services at IBM Consulting, the real opportunity isn’t just faster workflows—it’s rebuilding how business gets done from the ground up. In this episode, recorded in the heat of a New York summer, Francesco joined me to unpack what agentic AI really means for enterprise leaders and why “doing AI right” is about more than picking the latest model.

We began by breaking down the term agentic AI, which Francesco defines as the shift from passive assistants to intelligent agents that can actually execute work, not just suggest how to do it. That might sound subtle, but it’s a huge leap. And it’s not one companies can take by simply layering AI on top of broken or inefficient processes. Instead, IBM is helping its clients rethink entire workflows, starting not with the tech stack, but with the business outcome.

Francesco explains why data readiness is still the number one challenge. While many companies have talked about modernizing their data foundations, few have done it in a way that supports grounded, contextual, reliable AI agents. He introduces the idea of “data products” as a way to anchor agent behavior in the right context, feeding into IBM’s three-layer model: user experience, orchestration, and data.

We also explored the growing role of standards like Model Control Protocol (MCP), which could make secure integration with legacy systems more realistic at scale. Francesco highlights how IBM is addressing access control, security, and governance to ensure agentic systems are not only powerful but also trustworthy and accountable.

There’s plenty here for enterprise leaders wondering how to move AI projects out of pilot mode. From real examples in customer service, insurance, and pharma, to IBM’s internal strategies for employee upskilling, Francesco shares what early success looks like and why hackathons, hands-on experience, and human-centered design are critical.