Blog | May 20, 2026

The enterprises that win with AI will have the best foundation

Why getting to production fast is the most important move a CXO can make right now — and what must be in place to do it.

By Stan Clark, Vice President and General Manager of AI Market Development, DXC Technology

The technology is ready. The models are capable. Enterprise budgets are committed. And yet most AI initiatives are stalling before they reach production. The gap isn’t the AI. It’s the enterprise underneath it.

That’s not a criticism — it’s the reality of where we are. The enterprise foundations of data, integration, and process weren’t built for agents. Some organizations will reengineer what they have. Others will build new architecture alongside it. Both approaches work. The question every leader is wrestling with is how to move faster.

We’ve all relied on the same thing across every major technology wave — ERP, CRM, cloud. People. Smart people who knew the workarounds, filled the gaps and made imperfect systems hold together.

Agents don’t do that. They follow the system exactly as it’s built, which makes them the first honest X-ray of an enterprise’s operations. That’s not a problem. That’s a gift. For the first time, leaders can see precisely where the foundation needs work. The agent isn’t failing — it’s telling the truth.

That’s what makes this moment different from every technology wave before it. The ROI challenge isn’t about model capability. It’s about whether the enterprise underneath is ready for something that decides and acts on its own.

The foundation that must hold

In the production environments DXC operates — including core banking, claims platforms and government operations — three elements must work and build on each other.

Data. It must be reachable, connected and trusted. Siloed data means agents can’t reach it. Connected but ungoverned data means agents can’t trust it. Neither delivers value.

Integration. Agents live on APIs, coordinating work constantly across the business. Undocumented, inconsistent or fragile APIs stop agents cold. This is where Boomi becomes essential — not as an add-on to AI, but as the nervous system agents run on. A unified integration layer means agents can reach across SAP, Oracle, Salesforce and inherited systems without hitting each one directly.

Process. Processes must be clear enough to explain to someone who doesn’t know the workarounds. Agents don’t ask permission. They decide and they act. Enterprises can’t run agents on institutional knowledge that lives in someone’s head.

Get those three right and the same systems start producing different outcomes. The foundation stops being the constraint and starts being the competitive advantage.

Getting to production: where the real learning is

Most organizations design their AI initiatives for the pilot. Pilots have curated data, narrow scope, and eyes on every step. Production is a different environment — agents run unsupervised, handle exceptions nobody planned for and keep working around the clock.

The organizations moving fastest understand something important: the pilot teaches you what could happen. Production teaches you what does. That’s where the real learning is, and it’s why getting there quickly — with the right foundation — is the move that matters.

Speed, however, demands clarity about what enterprises put into production. When you deploy an agent, you’re not deploying software. You’re delegating judgment.

That distinction shapes everything about how an enterprise builds its operating model. Delegating judgment to an agent means the organization is accountable for what that agent decides and does. Governance isn’t a compliance exercise — it’s what makes it possible to move fast with confidence.

The leaders getting this right are answering four questions before agents go into production:

What are we letting it decide?
Who is accountable when it’s wrong?
How do we control it?
Who can stop it?

Answer those early and leaders can design the right operating model from day one. They also move faster — because legal, compliance and risk aren’t catching up to decisions that were already made.

The real prize

Cost and efficiency are the entry point. They’re not the destination.

When the foundation holds and agents can act across an integrated enterprise, the nature of what’s possible changes. Loans approved in minutes. Claims routed instantly. Customers helped while they’re still asking. That market where the math never worked before? It works now.

The best people in the organization stop processing transactions and start designing how the system runs. The question stops being “where can we cut 20%?” and becomes something more interesting: what can we create that we couldn’t create before?

That’s the shift worth moving toward, where AI stops being a cost play and starts becoming a growth engine. And the foundation — data, integration, process, governance — is what makes it reachable.

About the author

Stan Clark leads NextGen AI Partnerships & Growth at DXC Technology, driving co-innovation and partner-sourced growth across DXC’s strategic AI landscape. He has spent more than 20 years building enterprise partnership ecosystems across global systems integrators, hyperscalers, and ISVs.