Presented by Celonis
Oklahoma State discovered its blind spots the hard way. In April 2023, a legislative report revealed that its agencies had spent $3 billion without proper oversight. Janet Morrow, director of Oklahoma’s Risk, Assessment and Compliance Division, worked to track thousands of monthly transactions across dozens of disconnected systems.
Sooner State became the first US state to implement Process Intelligence (PI) technology for procurement oversight. Morrow says the change was immediate. Real-time monitoring replaced multi-year audit cycles. Market-leading Celonis’ platform immediately identified more than $10 million of improper spending. And the inspection team was able to redeploy staff from 13 to 5 members, dramatically increasing effectiveness.
“Process for Progress”: A Global Movement
Oklahoma’s pioneering success using powerful new process technology highlights an emerging global trend. Morrow was among more than 3,000 leaders who gathered at Celosphere, Celonis’s recent annual conference, to explore how AI driven by PI from a business context can deliver commercial returns as well as environmental and financial benefits around the world.
Vision: Treat intelligence as the foundation of public and social progress.
This movement sees the combination of AI and PI, like in Oklahoma, as a powerful way to help governments and other organizations provide critical services more cost-effectively, with better decisions and better-informed policies. From procurement to juvenile justice to health care and the environment, many organizations are now taking a first look at this famously Byzantine, opaque way of working.
For veteran financial leader Aubrey Vaughan – who is now vice president of strategy for the public sector at Celonis and was formerly a top executive at a major financial software firm – taking steps toward actual process improvement has taken a long time. He remembers proudly testifying before Congress a few years ago about uncovering $10 billion in improper government payments at his previous company. Later, a senior government official pulled him aside and suggested he downplay the achievement.
The reason, he was told: "The next question they’re going to ask you is, ‘Why is this happening?'” Vaughn says. “Today we can answer not only why, but how we fix it."
Across America and around the world, public agencies are facing tight budgets. The desire to deploy AI to bridge the gap is colliding with a difficult reality: You can’t automate what you don’t understand. Here are three real-world examples of organizations using PI and AI for better results.
Oklahoma: Real-time AI spending analysis increases accountability
Within just 60 days of implementation, Celonis reviewed $29.4 billion worth of purchase order lines, identified $8.48 billion in statutory exempt purchases and flagged problematic transactions. The system now provides real-time feedback to buyers within 15 minutes of purchase, allowing immediate improvements.
The system revealed that agencies were purchasing from a single vendor at a price 45% lower than the statewide contract, forcing renegotiations.
"Real-time AI analytics increases accountability by providing critical insight into spending patterns and streamlining contract utilization." Morrow explains.
Last year, Oklahoma adopted Celonis’ CoPilot feature, which uses conversational AI to let officers ask questions in simple language. Now, when the governor or Cabinet member has questions about a contract, they get an answer in seconds, not weeks, Morrow says. His group is expanding the technology to other agencies. It is also exploring how emerging AI agent capabilities can automate compliance and expense analysis.
In Texas, uncovering a surprising hidden pattern in young offenders
Erin Espinosa’s work at Evident Change, a social research nonprofit, is about good stewardship — not of taxpayer money, but of young lives.
Analyzing 400,000 data points from the juvenile justice and public health systems in Texas, the former probation officer turned Ph.D. Became. Made a startling discovery: The mental health treatment that young offenders received (or did not receive) was a stronger predictor of incarceration than the severity of the crime that brought them into the system. Espinosa told the courts, legislatures, Congress. Nobody believed it.
Frustrated, he partnered with Monica Chiarini Tremblay, a professor at the College of William & Mary. While traditional analysis showed correlation, Celonis’ process intelligence helped the pair show clear, quantifiable causation: a broken mental health system was actively pushing kids toward worse outcomes. Further machine learning analysis also showed that doubling the number of similar interventions increased the likelihood of unwanted out-of-home placement for juvenile offenders.
Recently accepted for academic publication, the real-world findings represent both an indictment and an opportunity. Espinosa and Tremblay are planning a larger 2026 pilot implementation of PI-based analytics by bringing together social services, juvenile justice, mental health providers, and education officials.
"It is a perfect intersection of business, social work, adolescent development and community financial implications." Espinosa says.
They are now exploring how AI agent technologies can flag at-risk youth and initiate coordinated responses before patterns become established.
The $1 trillion defense budget – which has never passed a clean audit
The US Defense Department is facing massive financial challenges. As acting Secretary of the Army, Robert M. Speer hired a big-three accounting firm to map out the service’s financial processes. After three years, the analysis was obsolete – the procedures had changed dramatically.
So, when Spear first saw process intelligence, he was really excited by what it revealed. "I can see not only the data,” he explained, “but also where it’s coming from, the business process that delivers it."
Tom Stephens, former deputy chief financial officer of defense, agrees: "Clearly a piece of the puzzle is missing." Both have recently joined Celonis’ Public Sector Advisory Board. They see the potential for AI agents to automate compliance monitoring across the DoD’s complex ecosystem.
The stakes are unimaginably high. The Defense Department will receive more than one trillion dollars in funding in fiscal year 2026. It is also the only federal Cabinet agency that has never passed a clean audit.
Beyond calculations, rapidly changing geopolitics and modern warfare demand dynamic systems just like current combat environments.
"We’re talking about the ability to make changes in real time," Spear says. "We know that’s what happens on the battlefield, but we need something behind those enabling processes and systems to make sure it happens correctly."
The pair are working with defense leaders to show how process intelligence can lay the foundation for change – enabling modeling and scenario planning that can support battlefield decisions with data-driven confidence rather than delayed, obsolete information.
Efforts to modernize and optimize complex government systems and processes have received a big boost recently. Working with partner Knox Systems, Celonis received FedRAMP authorization earlier this year, a security credential required for federal cloud services.
"Knox powers the most secure and longest-running managed federated cloud," Notes CEO Irina Denisenko, supporting 15+ federal agencies. Authority gives way to technology "As the backbone of compliance for the next generation of government SaaS."
Where process meets purpose
Early public sector adopters are proving what’s possible with process intelligence – from identifying billions in potential savings to revealing why children enter the prison pipeline. The possibilities also extend where public money shapes the public good: climate response, education, infrastructure, emergency services.
Advocates often speak of “process for progress” or "process for empathy" – Using transparency to change hearts and minds, not just policies.
Says Chiarini Tremblay, who worked on the Texas juvenile delinquent system: "We have to understand complex systems and make data-driven decisions, but the goal is always to improve outcomes for people."
This is not just an American movement. For example, in the UK, University Hospitals Coventry and Warwickshire NHS Trust have deployed PI with dramatic impact. Director Andy Hardy used Celonis to analyze 244,000 outpatient cases, revealing huge variation in care delivery.
By optimizing appointment reminders four to 14 days before visits, the Trust enabled earlier cancellations and saw an additional 1,800 patients weekly. The waiting list was reduced by 5,300 patients in eight weeks.
Hardy concludes: "For physicians, understandable data is as important as a scalpel."
Technology is continuously advancing. At Celosphere 2025, Celonis unveiled several new offerings and platform updates for public and private sector organizations, including an orchestration engine that coordinates tasks across workflows involving AI agents, human tasks, and legacy systems.
All Celonis are built on the Process Intelligence Graph, which creates a "living digital twin" The processes of a business or public agency. It is system-agnostic, working across disconnected systems typical for government operations – integrating together decades-old mainframes and cutting-edge cloud applications.
However, agency heads and others note that success requires more than software. For example, when Oklahoma reduced its inspection team from 13 to 5, resistance emerged. Morrow’s team invested heavily in training and change management. Process intelligence reveals opportunities for improvement, but people implement solutions’ she explains.
Continuous, long-term education and cultural change is needed.
“Continuous operational improvement is a lifestyle,” says Celonis Vaughan. “You need a culture that wants to create better processes, better systems, more efficient systems.”
The equipment is ready. The business case is proven. All that remains is the desire for change – and the courage to look clearly at the systems designed to serve the public good.
Sponsored articles are content produced by a company that is either paying for the post or that has a business relationship with VentureBeat, and they are always clearly marked. Contact for more information sales@venturebeat.com,
<a href