How Boomi unblocked 2,000 citizen developers to build AI agents with Tonic Structural

The challenge: A company-wide AI mandate with no safe way to test

Boomi is the glue between applications—middleware that moves data from point A to point B for businesses across industries, from financial services to automotive. With the launch of their agentic AI solution, Boomi saw an opportunity to do more than just connect applications: they could run data through intelligent workflows between those applications.

Then the CEO raised the stakes. "Our CEO wanted everyone in our company to become citizen developers, coming up with new ideas and solutions to help our business run on AI," explained Corey Webber, Director of Data Engineering at Boomi. "But those ideas needed data to be tested on. If somebody wanted to build an AI agent and a whole process around it, they didn't have a safe way of accessing data to run proof of concepts."

The challenge landed directly on Webber's team: "We need fake data, but we want it to be realistic because we need to build workflows that work just like they would in the real-world."

The problem was clear, but so was the gap. The team lacked a quick, reliable way to generate quality synthetic data. 

Without a solution, Boomi's citizen developer initiative risked stalling before it ever gained momentum. The company needed synthetic data that was realistic enough to power production-grade AI agents, generated fast enough to keep pace with the mandate, and secure enough that no customer data would be at risk.

The solution: From discovery to proof of concept in under six weeks

Boomi initially explored manual Python scripts and Snowflake-native options, but they soon realized these approaches wouldn't scale. After a recommendation from their CIO, the team turned to Tonic.ai. The requirement was strict: they needed a successful proof of concept in hand in less than six weeks—and that included the time spent finding the product.

Tonic Structural stood out immediately because of its intuitive UI and seamless integration with Snowflake. And the implementation was remarkably fast. “Tonic gave us what we were looking for at the speed that we were looking for,” Webber noted.

What struck Webber most was how little ramp-up was required. "I'm a big fan of good UI design. Without reading your documentation, without watching your tutorials, I was able to successfully configure a database and launch a workflow within two days."

A critical factor in choosing Structural was its configurability. Unlike other tools that felt rigid, the Tonic platform allowed the team to handle whatever complex data requirements arose, validating that the solution would work at scale.

“Everybody else we had looked at didn't have the configurability that we were looking for,” Webber shared. “Every time I said, 'I wonder if Tonic has this?', we were able to find it. Tonic gives us enough tools to make everybody happy."

Furthermore, the ability to refresh data on demand via API ensured that the synthetic environments remained current and useful for the development teams. Today, data refreshes run automatically over the weekend so up-to-date data is ready to go Monday morning.

The results: From citizen developer idea to enterprise AI agent with the flip of a switch

With de-identified data from Tonic Structural, Boomi has opened its citizen developer program to approximately 2,000 people. Dozens of employees are actively leveraging the synthetic data warehouse to build agentic workflows. By providing a secure, synthetic Snowflake environment, Boomi’s enterprise AI team has removed the risk of production data exposure while empowering the workforce to experiment freely.

The most powerful result is how smoothly solutions move from prototype to production. Employees build their AI agents using synthetic data from Tonic Structural, and when a solution is ready, the enterprise AI team simply points it at the real data. "Our employees build on the fake data, then our enterprise AI team implements their solutions. All we need to do is switch the URL for the Snowflake server from the synthetic data warehouse to the prod warehouse. And the solutions work. By just flipping a switch, we can activate these agentic workflows for the entire company. They just work."

The productivity gains are tangible. Currently, five high-intensity agents are being rolled out, some processing up to 60GB of data. These include:

• A sales insights agent that surfaces customer and account insights for sales calls, preparing tailored, targeted call scripts.
• A customer success agent that performs account analysis across data sources, surfacing retention insights immediately available on live calls—no digging required.

The success of the program has created a surge of internal demand. What started with the most actionable datasets is now expanding across all applications Boomi works with. The team is even working with training departments to expand their employees’ skills using the synthetic environment, further lowering the barrier to data literacy while scaling data access and AI innovation across the entire organization.

“Tonic has created this great success story internally,” Webber concluded. “It's broken down the barriers of getting people access to data. People are able to get the data they need, and security is on board.”