At Boomi World Forum 2026 in Singapore, Boomi unveiled an “agentic control plane” built on its active data foundation. Boomi CTO in APJ, David Irecki, also took the opportunity to reminisce about Boomi’s origins as an electronic data interchange (EDI) player that evolved into an integration-as-a-platform (iPaaS) player that specialised in API management.
We saw early on that there’d be data synchronisation problems across organisations.
David Irecki
“We saw early on that there’d be data synchronisation problems across organisations,” David explained, adding that the company had made a number of investments to bolster their data management capabilities and even “essentially put (our API management) in the top right quadrant in so many analyst reports.”
Humans and agents
Amy McNee, Boomi’s SVP of global solution engineering and data management GM, wanted to click into how problems with data are not new. But agentic AI is now expected to access more data and execute more tasks on behalf of humans, amplifying the data risk and its blast radius. Unlike humans, agents “reason on exactly what you’ve given it,” increasing the chance that poor or uncontextualised data produces harmful or misleading outputs.
“Companies have been kicking this problem down the road for a long time. They don’t have a choice (but to act) now because the exposure when an agent hits that data, it’s much bigger.” She even said that Boomi observed, “…companies taking from their AI budgets to clean up their data because data and AI are inextricably linked… now the data is the brain of the agent.”
Another trend observed is companies rolling back AI agents after discovering humans are cheaper and “could have stayed on as a traditional deterministic process.”
This is culminating into what David said is two categories and two processes – humans and agents, deterministic and probabilistic – mixing to achieve outcomes that the business needs.
Agentic control is essential and should take governance along
So, the plan is to control agent costs and risks by combining governance and architecture.
Boomi presents its Agent Control Plane as a vendor- and model-neutral infrastructure for governing agent activity, data access, cost and security rather than a standalone governance add-on. It works together with the AI Gateway which is more specifically the enforcement and traffic-control layer, with capabilities to detect, redact, and sanitise sensitive or injected content in prompts and responses.
…governance is now not a destination. It’s all part of the journey… attached to the data and to the agent.
Amy McNee
Built on technology from Boomi’s acquisition, Lunar.dev, this AI gateway can also enforce governance for MCP interactions. Through Boomi Connect, AI tools can reach enterprise applications through managed MCP services and a gateway with identity-based access controls and audit logging. This together with API management and an MCP registry enables governed interactions.
Capabilities like Knowledge Hub and MetaHub attach context, metadata, and governance directly to the data, reflecting Amy’s words that, “…governance is now not a destination. It’s all part of the journey… attached to the data and to the agent.”
Roadmap
Boomi is shipping quickly with products like MetaHub already generally available (GA). In summary, the company is anticipating customer needs like governance, AI gateway capabilities, and more before customers fully articulate them.
“Things like MCP service support, the AI gateway, MetaHub, Knowledge Hub, if you would have talked to somebody 12 months ago or 18 months ago, they might not have known they were going to need these things…
“I think everybody should be really excited about Boomi’s roadmap, and I know that I am. I think we have the vision, and we’re getting in front of it really fast, so we’re looking around corners, and we’re doing it right.”
