Behind the delegate pass of IBM THINK 2025

IBM think

United Kingdom, May 19, 2025

A New Era for Enterprise AI

Authored by Jo Walsh, IBM Solutions Director, Logicalis UK&I

At THINK 2025 in Boston, IBM didn’t just unveil new technologies it reframed the enterprise AI conversation. 

For CIOs, the message was unmistakable: the age of AI pilots and proof-of-concepts is over. The next wave is about operationalising AI at scale, with governance, interoperability, and measurable business value at the core.

Across six keynotes and dozens of breakout sessions, IBM executives, partners, and clients painted a picture of AI not as a moonshot, but as a maturing discipline—one that demands architectural discipline, cross-functional orchestration, and a relentless focus on outcomes.

AI Agents: From Concept to Core Capability

A standout announcement was the expansion of watsonx domain-specific agents, prebuilt for HR, procurement, customer service, and sales. These agents are designed to integrate natively with platforms like Workday, Salesforce, and Coupa, enabling CIOs to automate workflows without overhauling existing systems.

Not all AI is built the same, and not all AI is built for the enterprise. If you want to unlock the 99% of enterprise data that’s untouched by AI, you need models that are smaller, faster, and tailored to your business.

Arvind Krishna
CEO
IBM

Krishna’s point was clear: the future isn’t about chasing the largest model it’s about fit-for-purpose AI that can ingest enterprise data securely and deliver value quickly.

To support this, IBM introduced the watsonx Orchestrate Agent Builder, a no-code tool that allows business users to create and deploy agents with built-in observability and governance. This democratization of AI development is a game-changer for CIOs seeking to scale automation without bottlenecking central IT.

Data Fabric: The Bedrock of AI ROI

IBM’s enhancements to watsonx.data reflect a growing recognition that AI is only as good as the data that fuels it. The new unified data lakehouse architecture includes built-in lineage tracking, fine-grained access controls, and native support for open formats like Iceberg and Delta Lake.

Data is your true differentiator. But less than 1% of enterprise data is being used to train AI models. That’s the opportunity—and the challenge.

Rob Thomas
SVP of Software and Chief Commercial Officer
IBM

Thomas highlighted how IBM’s approach to hybrid data integration spanning on-premise, cloud, and edge—enables enterprises to break down silos and activate data wherever it resides. This is especially critical as AI workloads become more distributed and latency-sensitive.

Hybrid Integration for a Hybrid Reality

The launch of WebMethods Hybrid Integration, a result of IBM’s acquisition of Software AG’s integration portfolio, underscores a pragmatic truth: most enterprises are not “cloud-first” but “cloud-when-it-makes-sense.”

This new offering provides a unified platform to connect APIs, legacy systems, and cloud-native apps—reducing the friction that often stalls digital transformation. For CIOs, it’s a welcome acknowledgement that modernisation doesn’t mean abandoning what works.

We talk about 1 billion new applications being built over the next four years. Agents will form at least a third of them. But they can’t live in isolation—they need to connect to the systems that run your business today.

Arvind Krishna
CEO
IBM

Infrastructure That’s Secure, Sustainable, and AI-Ready

IBM also unveiled LinuxONE 5, its next-generation mainframe optimized for AI workloads. With support for confidential containers, quantum-safe encryption, and energy-efficient processing, LinuxONE 5 is designed to meet the dual demands of security and sustainability.

Generative AI is moving fast, but most infrastructure isn’t ready to support it. The gap between ambition and reality is widening—and CIOs need to close it before it becomes a liability.

Rob Thomas
SVP of Software and Chief Commercial Officer
IBM

IBM’s answer is a modular, composable infrastructure stack that supports AI at the edge, in the cloud, and on-prem, giving CIOs the flexibility to deploy where it makes the most sense.

Keynote Highlights: From the Racetrack to the Boardroom

The keynote stage at THINK 2025 wasn’t just about technology, it was about real-world impact. From Scuderia Ferrari HP to the UFC, IBM showcased how AI is transforming industries.

In a conversation with Ferrari’s Frédéric Vasseur, Krishna highlighted how watsonx is turning real-time race data into personalised fan experiences.

“AI is the productivity engine,” he said. “It unlocks the value in your data.”

Meanwhile, Lumen Technologies CEO Kate Johnson shared how IBM’s AI is helping optimise their fiber network and edge data centres:

“Every millisecond matters. It’s not quite like a Ferrari, but the same concept.”

These stories weren’t just flashy use cases—they were proof points that AI is delivering value today, not in some distant future.

What CIOs Should Take Away

THINK 2025 wasn’t about hype—it was about hard-won progress. IBM’s announcements reflect a maturing AI ecosystem where tools are modular, interoperable, and grounded in business value.

Here are five takeaways for CIOs:

  • AI agents are ready for prime time
    As long as they’re integrated into your existing workflows.
  • Data strategy is now AI strategy
    Invest in platforms that unify, govern, and activate your data.
  • Hybrid is the default
    Choose tools that bridge the old and new, rather than forcing a rip-and-replace approach.
  • Infrastructure matters
    AI won’t scale without secure, performant, and sustainable foundations.
  • Focus on orchestration, not experimentation
    The winners will be those who can align people, processes, and platforms around AI.

As Krishna put it:

“The hype cycle is fading. We’re now thinking about adoption, ROI, and business value.”

For CIOs, the challenge is no longer whether to adopt AI—it’s how to operationalise it at scale, with trust, transparency, and measurable impact.

 

Topic

Related Insights