Florida Crystals puts process intelligence at the heart of its operations

Florida Crystals, a pioneer in the US organic sugar business, is making process intelligence, powered by AI, the backbone of its operations.

Process intelligence combines process mining, task mining, and business intelligence to map, monitor, and optimize business workflows in real-time. Florida Crystals CIO Kevin Grayling prefers to think of process intelligence as advanced automation. He says the technology offers flexibility when it comes to operations, enabling the use of traditional AI, gen AI, or agentic AI for automation where it makes sense, and a reliance on less sophisticated automation where it doesn’t.

“It’s more flexible, so we’re able to go in and out of AI with process intelligence rather than exclusively out of process mining into AI,” Grayling says. “There’ll be many cases where process intelligence without AI is enough, but when you need more, it’s possible.”

Refining a complex business model

Florida Crystals, founded in 1960, is a fully integrated sugar cane company that farms sugarcane, rice, and vegetables on more than 190,000 acres in southern Florida. It also owns two sugar mills, a sugar refinery, a packaging and distribution center, a rice mill, and a renewable power plant that uses sugarcane fiber to generate energy. Grayling says Florida Crystals is really three companies: an agricultural company that grows sugarcane, a manufacturer that refines raw sugar from sugarcane, including sugarcane from other producers, and a consumer packaged goods business.

Grayling says layering process intelligence into its core systems enables the company to eliminate manual rework while scaling automation across multiple functions. In turn, he says, that helps to accelerate transformation and maintain enterprise scalability, security, and governance.

To begin, Florida Crystals focuses on leveraging process intelligence for its finance, procurement, and inbound supply chain operations. Grayling says the efforts should unlock millions of dollars in working capital, and help it automate processes around sales tax and state taxes.

“It minimizes incorrectly filed taxes and minimizes the time we have to spend with third parties to review taxes to make sure we’re not overpaying or underpaying taxes,” Grayling says.

Natural sweeteners meet artificial intelligence

Florida Crystals is also no stranger to AI. Its R&D team has been leveraging homegrown AI models for years, and sales, marketing, and HR have benefitted from RPA and ML. In addition, the company was an early adopter of Microsoft Copilot for its knowledge workers.

“Our plan this year is to significantly increase the agentic and generative AI capabilities in all of those functions,” Grayling says, adding that the steering committee considered 14 new use cases in areas including procurement, accounts payable, and customer service. “Probably about half of those are what you’d call true, neural learning AI rather than less sophisticated automation.”

In procurement, the company has developed a model that monitors plant maintenance, including indirect materials for safety stocks and ordering ahead.

“We implemented that a few months ago, and it significantly reduced our working capital in that area,” he says.

Other plans for procurement include automating requests for quotes from suppliers, analyzing quotes, and picking optimal options.

Strengthening partnerships

Florida Crystals works closely with process intelligence specialist Celonis as a partner to power its transformation. Specifically, Celonis is the company’s enterprise platform for AI and automation that enables orchestration of workflows across systems, teams, and technologies, and surfaces process intelligence insights within Microsoft Teams and through AI solutions deployed via Microsoft Copilot.

Roughly 95% of the structured business data the company creates is in SAP workflows and SAP data models.

“We anticipate that most partners will take time to fully enable their capabilities with AI,” Grayling says. “Maybe that’s a couple of years or longer in some areas. In the meantime, we’ll continue to use SAP process intelligence in a packaged not bespoke way to automate and add AI capabilities to certain business processes.”

Grayling’s advice to CIOs pursuing their own process intelligence efforts is to keep things as simple and streamlined as possible.

“You can be very consolidated and simplified,” he says. “You can be very software-as-a-service and you can apply that same logic to automation, including AI. That’s where our approach of embed first, process intelligence second, and bespoke third comes from.”