The news
Forrester's Adaptive Process Orchestration Software Landscape, Q2 2026 report marks a clear shift from task-level automation to full process orchestration at enterprise scale. At CamundaCon this week, Camunda's CEO Jakob Freund made a pointed observation: every process in your organization is legacy — because it was designed before AI existed. Their answer is blending adaptive AI behavior with deterministic workflows, not replacing structure entirely.
Our take
This is the conversation GTM teams need to be having, and almost none of them are.
Most demand gen and RevOps teams are trying to bolt AI agents onto processes that were never documented in the first place. The sequence looks like this: someone sees a demo of an AI SDR or an automated nurture workflow, gets excited, tries to implement it, and six weeks later it's either off or on life support. Not because the tool was bad. Because the underlying process it was supposed to automate was tribal knowledge living in someone's head.
Freund's framing — that every existing process is legacy — is sharp and practical. He's not being philosophical. He means that your lead routing logic, your MQL handoff, your follow-up sequence: none of it was designed with AI in the loop. Which means you can't just add an agent on top and call it done. You have to actually remodel the process with AI as a first-class participant.
The Forrester finding that resonates most is the emphasis on deterministic orchestration as the backbone. This is what works in production: AI handles the adaptive, judgment-heavy steps — personalization, prioritization, response generation — while the workflow itself is structured, auditable, and predictable. Human-in-the-loop isn't a fallback. It's an architectural decision made up front.
The teams that are winning with agentic AI aren't the ones with the most sophisticated models. They're the ones who did the unglamorous work of mapping their processes before they started automating them.
The so-what
The bottleneck in your AI rollout probably isn't the agent — it's that no one has written down how the process actually works today. Before you evaluate orchestration tools or scope an agentic workflow, get the current-state process out of people's heads and into a document. That work doesn't feel like AI implementation, but it is — and it's what separates the teams that ship from the teams that stall. The hardest part of agentic AI isn't the AI.
---