16. June 2026 · AI

AI in Enterprise Platforms 2026: What SAP, Microsoft, Salesforce, Oracle, ServiceNow and Workday Are All Announcing

Some years, the big software vendors use their conferences to set themselves apart. In 2026, the opposite happened. This spring I followed the six most important enterprise keynotes — SAP Sapphire, Microsoft Build, Salesforce Connections, Oracle AI World Tour, ServiceNow Knowledge and Workday DevCon — and the most striking thing was not what divides them. It was how strikingly similar their message has become.

This article is my overview of that season: what was actually announced, what the vendors have in common, and what you should take away as a decision-maker. It is the entry point to a short series — the strategic patterns, the required skills, and the real enterprise impact each get their own, deeper piece.

The short answer first

All six platforms are telling the same story in 2026: the systems where companies hold their data and processes are turning into acting agent systems. The industry consensus I heard on every stage, phrased differently each time, is this: from system of record to system of action. You no longer operate the software — it does the work. I described this business-model shift in detail in “From SaaS to GaaS”; here I line the platforms up side by side.

Six stages, one message

In fast-forward, what each vendor built its season around:

  • SAP Sapphire turned BTP into the Business AI Platform and unveiled an Autonomous Suite with, by its own count, 224 agents and 51 assistants across five business domains. The ERP, SAP argued, is “the brain” that gives agents context and governance. More in my Sapphire analysis.
  • Microsoft Build laid out a continuous AI stack from local hardware to the cloud — and introduced Agent 365, a control plane that gives agents identity, access and protection, even across AWS and Google.
  • Salesforce Connections declared the “Agentic Enterprise”: Agentforce agents as creative partners across every channel, grounded in your own customer data and secured by a Trust Layer.
  • Oracle AI World Tour nailed the principle “built in, not bolted on” — AI belongs in every layer of the stack, starting at the database, not as an add-on.
  • ServiceNow Knowledge positioned itself as the “AI Control Tower for business reinvention” and made governance the core product — including a kill switch for agents that go off the rails.
  • Workday DevCon opened the platform to external developer tools and introduced a Developer Agent that builds applications in natural language — with zero tolerance for error in payroll and the ledger.

The five shared patterns

Behind the product names lie five patterns that appear at all six vendors. I cover them in depth in “From System of Record to System of Action” — here is the short version:

  • Data and context are the moat. “The language model doesn’t know your business data” — that sentence was everywhere in spirit. The value lies in your proprietary context.
  • The semantic layer becomes mandatory. Ontologies and semantic models so the AI knows how to reason over your data.
  • Governance is architecture, not a feature. Every vendor built a dedicated control plane for agents.
  • Open protocols (MCP, A2A) everywhere. Agents cross vendor boundaries — and Anthropic’s Claude is a partner at five of the six.
  • Building agents goes low-code. Every platform now has a workbench to create agents in natural language.

What this means for decision-makers

I draw three conclusions for the leadership level. First, the bottleneck shifts from technology to people: the scarce skill is no longer configuring a platform, but orchestrating, grounding and governing agents on top of it. What that means in concrete roles I describe in “The New Enterprise Skills for 2026”.

Second, governance becomes a prerequisite, not a nice-to-have — without a control plane, agentic AI turns into shadow IT. Third, the real value today is unevenly distributed: early adopters show impressive effects, yet most companies cannot quantify it. I take that honest assessment apart in “Agentic AI in the Enterprise”.

Bottom line

2026 is not the year one platform left the others behind. It is the year they all converged on the same architecture for the coming decade — acting agents on trusted enterprise data, wrapped in governance. The vendors have delivered. The open question is who, in your organisation, will build, enable and supervise these systems.

That intersection of strategy, platform and enablement is exactly where I work with companies. If you want to translate the 2026 platform season into a concrete roadmap for your organisation, get in touch.