Many enterprise AI projects look great in a pilot but fail in production because they lack the data foundation, multi-model orchestration, and strict governance required to deliver true ROI across regulated industries and mission-critical environments. We hosted this hackathon to challenge builders to move beyond prototypes, and design solutions that can deliver measurable business outcomes.

The day was defined by high energy, deep technical focus, and a shared commitment to building real solutions to real problems. A distinguished panel of judges from The AI Exec Board [link] evaluated each use case and presentation based on commercial viability, functionality, and meaningful use of AI:

• Gus Kimble, Head of Global Strategy & GTM for Agentic Data Cloud at Google
• Laurie Pratt Ehrbar, CMO at Salesloft
• Michael Turcotte, CRO of MarketOne International

The participants used Kizen’s AgenticOS to build enterprise workflows across regulated industries, including Financial Services, Healthcare, Government, Supply Chain, and more. The winners leveraged a broad range of features within the Kizen platform to solve real business problems and presented workflows and use cases that can tangibly improve people's lives.

The Three Winning Use Cases

All the participants were encouraged to focus on clear scope definition, problem framing, and mapping solutions to existing enterprise workflows. After an intense five-hour sprint, three standout presentations rose to the top:

The Grand Prize: “Food Radar” by Disha Anil 

Disha Anil presented a use case that has the potential to significantly improve food safety in the US while streamlining time consuming and manual FDA processes.

The Problem: There is a critical time gap, averaging 59 days, between a food firm initiating a recall and the FDA officially classifying its risk level. This delay leaves the public dangerously unaware of potential product hazards.

The Solution: An AI-driven triage system that uses five interconnected workflows to fast-track risk assessment and notifications.

• Key Workflows:
  
  • Priority Assessment Agentic Workflow: Sources historical recall data from the FDA Open-Source API, and uses four specialized agents (Historical Data, Distribution Impact, Compliance, and Orchestrator) to generate an immediate risk factor and reasoning memo. A review task will be triggered, and a human reviewer will double-check the AI reasoning. 
  • Complaint Early Warning Workflow: Monitors and clusters consumer complaints within a specific time period, and auto-creates a recall record and triggers a priority assessment agentic workflow if volume exceeds certain guardrails. 
  • Traceability Notification Workflow: Automatically flags and logs potentially high-risk recalls for individuals to review, before automatically notifying affected distributors and retailers in the supply chain. 
  • Review Approval Sync Workflow: Checks if the reviewer approves the AI classification.
  • Real Send Demo: SMS-notification to the human reviewer.

Judges' Comments: The judges praised the commercial relevance, noting that this use case and solution solves a massive, real-world food safety problem for governments and private sectors alike. They were particularly impressed by the many interconnected agents and workflows working together to solve a problem that most consumers are touched by today. Notably, Disha leveraged the entire Kizen platform and built a commercially viable solution in under five hours. 

Best Commercially Viable Solution: “Dora” or Denial Optimization & Recovery Assistant by Michael McEntee 

The Problem: Revenue cycle analysts in healthcare are overwhelmed by the manual labor of reviewing, analyzing, and prioritizing medical claim denials returned by insurers, which increases operational overhead and manual paperwork, and delays critical patient care.

The Solution: Michael used Kizen’s AgenticOS to build an AI assistant designed to automate the intake, analysis, and triage of these denials.

• Key Workflows:
  
  • Automated Claims Workflow: Parses claim amounts, appeal deadlines, and reason codes while evaluating supporting medical documentation.
  • Priority Metric Scoring: Cross-references deadlines with LLM reasoning for a probabilistic "likelihood of recovery" calculation, allowing analysts to prioritize high-value, actionable cases.
  • Human-in-the-Loop Rationale: Packages the claims with recommended actions and reasoning memos, queueing them for human review rather than altering billing data autonomously.

Judges' Comments:
The judges highlighted Michael’s thoughtful approach, specifically how he built a workflow with a real person and their challenges in mind. The interactions between the human analyst and the AI agent create learning patterns that continuously improve the solution. 

Best Use of AI: “Policy Overlay” by Saurav Das 

The Problem: As a former government contractor, Saurav frequently struggled to gain access to software, services, and tools he needed in order to perform job-related tasks. 

‍The Solution: Saurav used Kizen’s AgenticOS to build a “policy overlay” that establishes rules that determine how an employee or contractor can get access to various IT resources such as software, services and tools. When requests are generated, Kizen’s system will automatically process them.  

Judges’ Comments: The judges praised Saurav’s focus on the learning aspect of AI, noting that the solution observes user behavior to continuously improve. The solution Saurav built leverages the full capabilities of the Kizen platform while serving multiple business domains, including OCFO, OCTO, OCIO, and executive business lines in their processes to get a holistic view of IT costs, licensing, and users’ actual IT needs. 

At Kizen, we believe that the best of AI is about building intelligent systems and solutions that understand the complexities of highly regulated industries and mission-critical processes. Watching these builders translate operational problems into production-ready workflows in under five hours confirmed that the future of agentic orchestration is already here. Their abilities to leverage AI to build real solutions to real-world problems across regulated industries keep us all optimistic that AI can truly help people and organizations work better, and not just faster. 

Congratulations to our winners and to every participant who stepped up to build the future of work. 

We can't wait to see what you build next.

‍