Agentic AI is developing at a rapid clip with cloud giants, software companies and enterprises racing to create autonomous agents. The catch is that there little integration across platforms, standard are lacking and orchestration and context is missing.
That's the gist AI agent state of union via Ed Macosky, Chief Product & Technology Officer at Boomi. Boomi plans on launching an AI agent control tower March 10 and aims to address some of the looming issues for enterprises.
At Constellation Research's AI Forum in Silicon Valley, Macosky said managing AI agents will be seen as an "integration problem" with a lot of moving parts with process and the symbiotic relationship between humans and AI.
Boomi, which has partnered with the likes of AWS and ServiceNow, has bet on AI agents and its role in herding them into something coherent to drive ROI for enterprises.
- Boomi acquires Rivery, adds data integration tools to iPaaS stack
- Boomi CEO Lucas: AI agents will outnumber your human employees soon
- Boomi's Transformation: From iPaaS to Intelligent Integration & Automation
- Why enterprise, process workflows are the new battleground
Here's a look at a few agentic AI issues that need to be resolved:
Agent overload and fragmentation. AI agents are being launched in a best-of-breed approach and the technology is accelerating. "I spent a lot of time thinking about how these agents will work together. We don't want to get in the way of how agents will communicate and work with each other, but every vendor is coming up with their own protocols, their own standards," said Macosky.
He added that most CIOs don't know how many agents they have running in the enterprise. "Do you have visibility? Nobody has that answer today. The number one concern in terms of agents and AI and business is security, compliance, governance and risk," said Macosky.
Orchestration. Agents will have to work together on tasks and automate across business processes, but orchestration layers need to be built. "We're all in on agents being the future," said Macosky. "But the orchestration layer needs to bring them together into something intelligent."
How enterprises are building agents. Given Boomi's platform, Macosky has a view into how customers are building AI agents. "When you talk to different customers, some have an affinity to a hyperscaler they're working with. Some are using application vendors. Or it's all of the above," said Macosky. "There are multiple levels and technical ways of doing it bringing agents together."
He added that agents are even being built by citizen developers with low code tools.
- AI agents bring consumption models to SaaS: Goldilocks or headache?
- Agentic AI: Three themes to watch for 2025
- Agentic AI without process optimization, orchestration will flop
Use cases. Macosky said that customer service is the use case most likely to be in production, but HR is going to be one of the leading use cases. Finance is also a key area as a way to optimize processes. "Agents in HR might be the most like the most important agents in a business in the future," he said.
The need for context. Macosky said situational awareness will be critical to orchestrating humans. He said:
"We can keep adding context, but when you start getting the situational awareness, that's where human decisions and human emotion and backgrounds of people come into play. Technology wise, we can be there quickly. But once the agents are in place they have to learn and understand human behaviors. It will be more about time than technology."
Standards. Macosky said standards will be critical. "We're defining and working with the likes of ServiceNow, AWS, Google and others. Most of the systems integrators out there are contributing with us. It's a mix of standards and some centralized governance where we're going to be announcing," he said.
The need for standards and a neutral control tower is necessary because vendors are building AI agent management consoles, but only in the context of their platform.
