Why Enterprise AI Value Starts Below the AI Layer

Most enterprise AI conversations start at the top of the stack: agents, copilots, use cases. But the organisations actually creating value are focusing somewhere else. At Agentforce World Tour in Utrecht this week, that showed up repeatedly. The real work is happening below the AI layer.

What the room was actually paying attention to

One example of that pattern was E.ON One, the IT and digital services arm of one of Europe’s largest energy utilities. They moved from fragmented sales data and manual pricing to a unified quote-to-cash platform in under four months. Not a pilot or a concept, but a working commercial foundation with integrated pricing, consistent data, and real-time visibility into TCV, MRR, CLTV and margin.

Built on Salesforce Revenue Cloud and fully integrated with SAP invoicing, it supports complex subscription and hybrid revenue models. This is a foundation that agents can actually operate on.

The room was full because enterprise leaders recognise that story. They are not chasing AI capability. They are trying to understand what needs to exist underneath it. This is what AI readiness looks like in practice.

The pattern behind it

This is not just an implementation story. It reflects a broader pattern in enterprise systems. Economic value does not accumulate where innovation happens; it accumulates where systems are stable, governed, and deeply integrated. That is the core idea behind the Value Gravity™ framework.

At the top of the stack, AI capability moves fast. Copilots, agents, and predictive models are where innovation happens — but also where value commoditises quickly. In the middle, orchestration connects systems through journeys, workflows, and automation. Value builds here, but also erodes faster than most organisations expect. At the base, the commercial foundation — CRM, CDP, identity, data governance, consent — is where value accretes slowly and durably. The economic gravity pulls downward.

The pattern at Agentforce Utrecht was consistent across sessions. Shell’s agentic journey framework was built on governance guardrails before capability leaps. MuleSoft framed API-led integration as the prerequisite for the agentic enterprise. Presenters consistently distinguished between organisations trying to adopt agents as a bolt-on versus those that had done the foundation work first.

Three signals for enterprise leaders in 2026–2028

Platforms become AI’s orchestration spine. Enterprise systems are not being replaced by agents. They are evolving into the stable, governed layer that agents plug into. The organisations that will lead are those that own the context plane — the governed data and workflow layer underneath the AI capability — not just the tools at the top.

Governance shifts from cost to competitive advantage. With EU AI Act timelines solidifying, proactive data, identity, and policy governance is becoming a speed-and-risk advantage, particularly in regulated B2B sectors. The organisations treating governance as a sunk cost today are the ones who will face remediation costs when compliance requirements arrive.

Ecosystem depth beats isolated features. Hybrid stacks win when orchestrated well. A great copilot means little without deep integration into your revenue architecture. Point-solution AI tools that cannot connect to governed data and execution layers will not deliver at enterprise scale.

Early signals are visible across the ecosystem. Adobe’s platform evolution, Salesforce’s agentic push, and governance lessons from incidents like the Odido Salesforce breach all point in the same direction: value concentrates where defensibility lives. The organisations that are building the foundation now are the ones who will be positioned to use AI capability effectively as it matures.