376 employees signed up. 46 teams across the globe. 38 departments represented on 180 working submissions. Not pitches. Not slide decks. Working builds.
Those 376 came from 45 different cost centers: Sales, Legal, Finance, Customer Success, Support, HR, Marketing, and Engineering all represented. Roughly three in four were in customer-facing, not technical, roles. They were the people who live inside the broken processes every day. And 40% of them didn’t wait for a teammate. They just built something on their own.
Workato’s fourth annual hackathon looked different this year. Previous rounds focused on automation. This one opened the floor to AI, with every team building on Enterprise MCP servers connected to real business systems. The rules were simple: no production accounts, no free trials, no restrictions on who could participate or what they could build.
Here is what came back.
The CIO Who Built Alongside His Team
The highest-scoring project was an AI career coach built on Workato’s Agent Studio. The team: Jessalyn Klein, Effie Tan, and Jia Ying Lee, with CIO Carter Busse working alongside them.
They connected Workday, Confluence, and internal job boards through MCP, then built an agent that could match employees to open roles, surface relevant learning paths, and flag career moves they would not have found on their own.
A CIO could have judged this hackathon from the sidelines. Carter entered it. That says something about how Workato thinks about this technology internally: not a tool to evaluate from a distance, but infrastructure to build on.
A Legal Team That Built Its Own AI Engine
Nova won the Customer Impact Award. Allan Tan and his team connected LinkSquares, Jira, Confluence, and Workday into a custom MCP server architecture that unified a fragmented legal stack.
Before Nova, discovery and compliance reporting meant combing through contracts in one system, cross-referencing policies in another, pulling people data from a third. The kind of three-system, four-handoff process that never makes it onto a transformation roadmap because everyone just accepts it.
Nova turned that into a single query. Complex multi-step discovery, redlining, compliance reporting, all running through one AI-powered core.
Today’s legal teams are not slowed down by lack of intelligence. They are slowed down by data fragmentation. Nova fixed that in weeks, not quarters.
Support Tickets From 30 Minutes to Five
The Innovation Award went to a team in Japan: Senin Yamamura, Takuto K., and Riki Akahori. They built an advanced support automation engine using custom MCP servers and multi-genie orchestration working together.
The result: support tickets that previously took 30 minutes to resolve now take five.
PRDs That Write Their Own Engineering Tickets
Robert Fujara and Jasper Madrone won Best Use of AI with a system that reads Confluence PRDs and automatically generates Jira epics and stories, complete with a Jira integration and terminal coding.
Product documentation to engineering execution, bridged by an agent. No manual translation. No lost context between what product wrote and what engineering built.
The Pattern Across 180 Submissions
Across all 180 builds, a few things stood out.
The winning projects did not come from one department. They came from legal, engineering, support, HR, and product. The problems they solved were not on any IT backlog. They were buried in daily friction that only the people doing the work could see.
A contract review process with three handoffs. A support queue that burned 25 extra minutes per ticket. A career development path that lived in someone’s head instead of a system.
These are not the problems that make it onto a transformation roadmap. They are the ones that cost thousands of hours a year.
Across all 242 registered teams, the use cases mapped almost perfectly to where friction actually lives.
Productivity and automation topped the list:
- 43 teams built something to eliminate a recurring manual process.
- Sales intelligence came second, with 37 teams building account research, competitive intel, and Gong-powered coaching tools.
- Customer success, HR, product, and marketing all followed.
- No department sat it out.
Twenty-one percent of teams crossed departmental lines entirely:
- The problems that live at the handoff between teams, the ones that fall through because no single department owns them, finally had builders on both sides of the table.
- Sales paired with Marketing, Professional Services with Customer Success, Engineering with Product.
- Twelve teams spanned three or more departments.
“Four years of hackathons, and this was the first time the winning projects came from every corner of the company,” said Stephanie Dwight, VP, Business Technology, who has led Workato’s hackathon since its first year.
“Enterprise MCP servers changed who could build. Before, you needed to know how to wire up integrations. This year, teams just connected Claude to Salesforce, Jira, Workday, whatever system their work actually lived in, and started solving problems. People who had never built anything before were submitting working solutions.”
What IT Did Not Have to Do
No custom integrations. No six-month scoping projects. No dedicated engineering resources assigned to each team. IT made one change to deploy Enterprise MCP, and 46 teams ran with it.
Every project used the same pre-built MCP servers and the same governance layer. IT kept full visibility and control. The ideas, the builds, and the impact came from the people closest to the problems.
Your Builders Are Already There
Every enterprise has these people. The ones who know exactly what is broken and have ideas for how to fix it. They have never been empowered to build the solution because the tools required engineering support and the timeline was always months.
But having the right people is only half of it. You also need the connective layer that lets them act. Enterprise MCP is what gave 46 teams the ability to plug AI into real business systems, build real solutions, and ship them without waiting in an IT queue. The ideas were already there. The infrastructure to act on them was not.
You do not need every employee to become a builder. You need the five or ten or twenty who already want to be, and you need to give them the platform to do it.