Boomi builds a role for agents and guardrails in the data-connected enterprise

Artificial intelligence and machine learning have been a central focus for Boomi LP over much of its 26 years as a data activation company.

The organization began storing anonymized metadata from customer integrations when it launched its multi-tenant platform for the cloud in 2005. In 2010, Boomi added a wizard to its platform that relied on machine learning to leverage integration mapping patterns.

This paved the way for a series of AI-based offerings that Boomi has deployed to support over 30,000 customers globally, including more than a quarter of Fortune 500 companies. AI agents have become a major part of the firm’s business with more than 75,000 of them running on its platform. It’s all part of Boomi’s data-connected vision, according to Chair and Chief Executive Officer Steve Lucas.

“We came in 20 years ago, and we built the connected enterprise,” Lucas said in a recent conversation with theCUBE, SiliconANGLE Media’s livestreaming studio. “Over time, that evolved into the automated enterprise. What’s next is going from connected to automated to agentic, and that’s what we’re focused on.”

This feature is part of SiliconANGLE Media’s exploration of the architectural shifts powering continuous, production-grade AI. Be sure to check out SiliconANGLE’s extensive coverage of Boomi World 2026, airing May 13–14, featuring interviews with Boomi executives and other industry leaders. (* Disclosure below.)

Enabling agents with data activation

At Boomi World in Denver two years ago, Lucas outlined a vision of the agentic AI future. The company unveiled a number of new platform capabilities at the event that year, including AI agents for protecting personally identifiable information, autonomously documenting and building data integrations, and providing proscriptive help for data activation.

“Applications are going to get rewritten on this new technology, and Boomi absolutely must be a part of it,” Lucas told reporters during the 2024 event.

It proved to be a prescient move on Boomi’s part. AI agents have boomed in popularity since then, with a projected 44% annual growth rate between now and 2030.

Last year, Boomi rolled out Agentstudio, an AI agent lifecycle management platform designed to provide a no-code interface for building, managing and governing autonomous tasks. Last month, the company unveiled Meta Hub for establishing a shared source of truth to be used by agents and humans alike. Boomi also installed new governance for agents within its Control Tower offering and introduced intelligent agent integration guidance tailored to specific environments.

In an interview with SiliconANGLE at April’s HumanX AI conference in San Francisco, Lucas was asked why he was so certain in spring 2024 that agents would become a major tool for enterprises.

“I don’t think I’ve ever been more certain about a topic, even two years ago,” Lucas recalled. “I started to think a lot about the inflexible processes that all businesses have. I thought this may be the technology that bends without breaking.”

Containerizing the agentic risk

Bending without breaking involves an ability to use agents safely and securely within the enterprise. This has proven to be a challenge as agents have moved from answering questions to executing specific tasks. An example of this can be seen in the meteoric adoption of OpenClaw, an agentic artificial intelligence framework designed to run continuously and act on behalf of users. OpenClaw has rapidly grown to 3.2 million monthly active users, and it was featured prominently in the announcements from AI powerhouse Nvidia Corp. last month, which deployed its own version called NemoClaw.

Yet the NemoClaw announcement was accompanied by an admission from Nvidia CEO Jensen Huang that OpenClaw needed additional security controls, which his company also announced. Vulnerabilities in OpenClaw deployments have exposed tens of thousands of internet-facing instances to takeover in recent months.

“The reality is we’ve got to step back for a minute,” Lucas told theCUBE. “Jensen gets on stage and asks everybody, ‘Do you have an OpenClaw strategy?’ But the reality is the CIO, the CTO still have to make this stuff run in the enterprise. What we’re focused on end-to-end is how to create containers for all of these different types of agents.”

Lucas and Boomi make the point that agents should run with layered guardrails, established rules and in isolated containerized environments during the proof-of-concept stage. This approach is part of an evolving dialogue that is likely to involve government agencies at some point. The National Conference of State Legislatures has reported that all 50 states in the U.S. have introduced legislation on the topic of AI.

Building public/private partnerships to manage the growing deployment of AI will be important going forward, according to Songyee Yoon, founder of Principal Venture Partners and a member of the Board of Trustees at MIT. “Companies need guidance, we are missing that dialogue,” said Yoon, in an interview with SiliconANGLE during HumanX. “Guardrails enable us to be more creative and innovative.”

Cost issues with AI deployment

In addition to security, Boomi and other major firms in the AI arena are also dealing with the issues of cost and ROI. As recently noted by theCUBE Research’s analysts, the data suggests that although virtually all firms are adopting AI, those realizing return on investment at large scale remain the mid-to-low teens.

Survey data supplied by Enterprise Technology Research shows that a meaningful slice of organizations are building, experimenting and learning — but not getting payback yet. And the spending numbers are sizable. One report documented that annual enterprise spending on AI surged from $1.7 billion in 2023 to $37 billion in 2025.

“I’ll tell you what I am hearing from COOs, CIOs, CTOs, even CEOs,” Lucas told theCUBE. “They’re starting to get these bills from Anthropic, from OpenAI. And what I’m hearing is, ‘I’m spending 10x this year what I did last year.’ The productivity gain is fair, but what we’re also hearing is a lot of concern. Is this 10x a year going to continue?”

Boomi itself has sought to counter the current narrative on AI costs by focusing on the role of automation in contributing to key areas of business impact. These include an average of 260 hours annually in productivity savings, 50% improvement in lost revenue recovery through enhanced data visibility, and retirement by IT departments of smaller integration platforms.

Careful management of token and model usage will also help, according to Lucas.

“’How is the weather’ is a prompt,” Lucas noted. “Those tokens cost just the same as, ‘Write code for me.’ A token is a token at the end of the day. So, routing it correctly to the least expensive model gives you the best return bang for the buck.”

Building partnerships and making acquisitions

As it navigates the shifting currents around AI, Boomi has made a series of acquisitions and partnerships to bolster its market position. The company’s purchase of Rivery Technologies Ltd. in 2024 laid the groundwork for a suite of new offerings aimed at streamlining data integration and pipeline management. The acquisition of Thru Inc. last May expanded Boomi’s file-based integration capabilities within a single, cloud-native platform to manage data movement across APIs, applications and files.

Boomi has partnered with ServiceNow Inc. to reduce manual work through intelligent automation. Boomi employs ServiceNow’s App Engine to build workflows into Boomi Data Hub, its cloud-native data management platform, for a simplified and intuitive user experience.

The company has also formed a multi-year strategic collaboration with Amazon Web Services Inc. to help customers build, manage, monitor and govern AI agents across enterprise operations. The integration of Amazon Bedrock with Boomi Control Tower enables a centralized management solution for governing AI agents across hybrid and multicloud environments.

In October, Boomi announced a native Change Data Capture ingestion with SAP SE. The new service enables enterprise teams to securely ingest data from SAP ECC, S/4HANA and SAP Business Warehouse into any cloud data lake or warehouse without writing a single line of code.

These acquisitions and partnerships highlight the strategic approach that Boomi is taking to define its future role in enterprise AI. Businesses need a platform that can support agentic AI across a multitude of environments, and Boomi has charted its course to be a major player in this area.

“As gen AI changes the roles of developers, operators and business leaders, we see platforms that have the ‘DNA’ of applications on top of cloud-native technologies being the new Agentic Infrastructure leaders,” said theCUBE Research Executive Analyst John Furrier. “Boomi has the pieces to be a scalable platform to accelerate the surge in domain expertise in applications and business users.”

Here’s the complete video interview with Boomi’s Steve Lucas:

(* Disclosure: TheCUBE is a paid media partner for Boomi World. Neither Boomi, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Image: SiliconANGLE/ChatGPT