Broadcom and Enterprise Management Associates (EMA) ran a joint webinar on 6 July 2026 called Making Agentic AI Operational, and its central claim is governed orchestration: embedding AI-driven decisions inside workflows that are already auditable and policy-bound. The framing inverts the usual pitch, because it treats the barrier to enterprise AI as operational trust rather than model capability. The session is now available on demand, and it is worth an hour whether or not V26 is already on your roadmap.
What governed orchestration actually means
The session pairs Dan Twing, President and COO at EMA, with Automic product manager Kaj Wierda, drawing on EMA research and the Automic V26 release. The through-line is unglamorous and correct: most organisations already orchestrate ITOps scheduling, DevOps pipelines and DataOps/MLOps workflows across mainframe, distributed and hybrid environments. The safe route into AI is therefore to extend that existing orchestration, governance and execution over AI agents and APIs, not to stand up a parallel stack beside it. Broadcom’s own phrasing is that workload automation is evolving from scheduling into a broader enterprise control layer.
Why it matters to your Automic V26 decision
If Automic V26 is on your roadmap, this is the strategic case for it. Agentic AI, APIs and orchestration are converging, and governed orchestration is how you keep human oversight, auditability and rollback while letting AI take on real work.
In V26 that principle has a concrete shape. AI-driven decision-making is wrapped in the role-based access controls, audit trail and security protocols the platform already enforces, so an autonomous remediation step answers to the same rules as any scheduled job. There is no separate regime to police; the existing controls simply extend to a new object type.
5 questions to answer before you switch it on
The direction of travel here is clear enough. What varies is how fast any given estate can follow it. A platform can offer governed AI execution on day one; whether your team can approve, audit and roll back an autonomous action on day one depends on your policies rather than your release version.
EMA’s research is blunt about why that matters. Roughly 30% of organisations report hitting incorrect or problematic AI outcomes frequently, and more than 75% have already had to correct or roll back an AI-driven action. Worth settling first:
- Where in your workflows would you actually trust an AI-driven step today?
- What evidence would you need before widening that trust?
- Who approves an agent’s remediation at 03:00, and how is that approval recorded?
- What does rollback look like when the decision was non-deterministic?
- Which of your existing audit obligations already cover this, and which do not?
Governance is only as strong as the policy behind it, and the policy is yours to write. Broadcom’s follow-up analysis, Defining the Enterprise Control Plane, is a useful second read once you have your own answers.
How Tricise helps
We help Automic teams turn strategy into something operable. Our Automic Automation services and governance consulting define the policy, RBAC and human-in-the-loop model this approach depends on. Because V26 introduces genuinely new concepts, Automic training at Tricise University gets your operators and schedulers fluent before go-live rather than after an incident. If a V26 upgrade is already scheduled, we fold the governance work into the same plan as part of our Smart Automation approach.
Frequently asked questions
Is the Broadcom and EMA webinar recording still available?
Yes. The session aired on 6 July 2026 and is now listed as on-demand video on Broadcom’s automation site, accessible through their BrightTALK channel. Registration is still required to watch it.
What does governed orchestration actually change in Automic V26?
AI decision-making is encapsulated in a job type that inherits the platform’s existing role-based access controls, audit trail and security protocols. The practical effect is that an agentic step is governed by the same mechanisms as a scheduled job, rather than by a separate AI-specific control regime.
Do we need to upgrade to V26 before working on AI governance?
No, and waiting for the upgrade is usually the wrong sequence. Deciding who approves an autonomous action, what must be auditable, and where rollback sits is version-independent policy work, and it is the harder half of the job. The upgrade then implements a decision you have already made.
Does governed orchestration only apply to Broadcom environments?
No. The pattern of encapsulating AI decisions in an auditable object that inherits existing access controls is emerging across the orchestration market. V26 is the implementation most Automic teams will meet first, but write the governance model to be portable — most estates run more than one scheduler.
Where should an Automic team start with agentic workflows?
Pick one workflow with a small blast radius, define an explicit human approval gate, and run it under full audit for a quarter before widening scope. The evidence from that quarter is what justifies the next step, and it is far more persuasive internally than a business case built on projections.
Talk to us about V26 governance
Planning for Automic V26 and agentic workflows? Talk to Tricise about governance and enablement. One conversation, no obligation.
