The challenge: “the vibe is great, the governance is not”
In March 2026, the IT and security team at Tines grew concerned about employees building vibe-coded apps, with no visibility into where they ran or what data they touched. Teams like IT and security used Tines' original product, Tines Stories, to build workflows — but less technical teams like finance and marketing had started asking about access to tools like Claude Code and Codex.
Brandon Maxwell, Head of IT Operations & Information Security, sent an internal memo asking employees not to ship these tools without contacting his team first.
He pressed a simple point: working and production-ready are two very different things, and the gap between them created a huge amount of work for his team. "We have zero capacity to absorb this work right now," he wrote. His closing ask to employees was blunt: “If this broke, was breached, or disappeared tomorrow, would anyone know and would anyone be ready to deal with it? If the answer is no, please come talk to us first.”
Maxwell had no idea that in four months, everyone at Tines would be building their own software with AI. And that he and his team would be ready for it.
At the same time as Maxwell’s memo landed, teams across Tines were trying, with varying degrees of success, to build solutions to their own problems. One of them was Director of Finance Peter Coleman, who had spent years trying to solve a problem that quietly ate up his team's time every month: manually assembling billing data from Salesforce and NetSuite for every single deal.
He tried to build the fix himself with our original product, Tines Stories, and couldn't — his technical expertise and data access didn’t allow it. So he waited.
The solution: a complete environment for enterprise AI
In July 2026, every employee at Tines was given access to Tines 3B, a complete environment where anyone can build apps, agents, and automations using natural language, while IT and security retain visibility, governance, and control over that work.
The app sprawl Maxwell had warned about became a single, governed surface of thousands of builds, instead of dozens or hundreds of untracked and problematic ones. "We know the credentials, we know the data is encrypted and treated the right way," he says. "The first step of securing any environment is inventorying it and knowing what's there."
Having everything centralized in 3B really helps — we understand what connectors are connected to where, how data is stored, and when applications are deployed, we can enforce authentication and have incident response capabilities on top of it.
Brandon Maxwell, Head of IT Operations & Information Security
Some of that governance came directly from Maxwell's own team using 3B to solve their AI governance problems. Rolling out access management at scale raised a basic question: who actually has access to what connector, and why?
Maxwell built an access map himself, inside a 3B story, using the 3B API. "I posted it in a Slack channel and shared that this was really helpful for me, and that it'd be great if there's something like this in the product." The product team listened, and it became the popular feature now known as the access map.
Meanwhile, Peter Coleman was ready to start building. On the day everyone got access to 3B, he started working on the billing problems he’d been thinking about for years. Within a few weeks, his billing platform was live. Today, his whole team relies on it, along with several other core operating platforms built with 3B. None of them can imagine going back to how they worked before.
The impact: an organization of builders, no longer waiting for permission
For Martin Moroney, Tines' Head of AI Transformation, this growth was exciting, and faster than he'd ever expected. "It took our teams eight years to build 1,000 workflows with our original product, Tines Stories," he says. "We reached that in less than 90 days with Tines 3B, and those numbers have just kept rising."
90% of Tines employees now use 3B as part of their job, and every employee interacts with something built in 3B daily. Four months in, around 500 employees have created more than 3,000 builds, many of them experiments. But around 400 of those builds are now relied upon for key business operations.
The shift wasn't just in volume, but in who was building. "Before 3B, the top automating teams were the more traditional ones — engineering, customer success engineering, IT and security," Martin says. "With 3B, that's changed pretty radically. We see finance and marketing up there among the top teams."
The biggest surprise, Moroney says, was how quickly employees lost patience with vendor-supplied software and started looking to replace it. "They immediately started to think, ‘The tools I have today don't do a good enough job. So I'm going to build a new platform where I'll actually do my work.’"
That enthusiasm comes with an increase in AI token costs, but at Tines, this has remained at a manageable level. "In general, the amount of money organizations are spending on AI tokens is completely unsustainable," Moroney says. “That's one reason why 3B guides the user toward deterministic builds wherever it makes sense.” In fact, just 23% of the things Tines employees have built include an AI step.
Some of our most important builds at Tines are deterministic — meaning once they're built, they should be cheap to run forever.
Martin Moroney, Head of AI Transformation
Today, teams across Tines have replaced complex vendor-supplied software, from ITSMs to learning management systems to customer intelligence platforms, with 3B builds. Not primarily to save money, though dozens of these builds now save $50,000 or $100,000 each per year in software fees. The real reason is simpler: those other tools couldn't do the job teams actually needed them to do.
Moroney sees a bigger shift happening in how companies think about buying software. He points to Tines' revenue operations team, which was using a Salesforce dashboard that recently got a long-requested upgrade: going from three filters to five. "A RevOps team member can not only create that dashboard in 3B really quickly, but make it orders of magnitude better than the old one."
Maxwell also views the rollout as a success but doesn't claim that 3B removes all risk associated with AI. It just makes the risk manageable. "You can think of it like taking a flight,” he says. “I'm not 100% certain about the outcome, but I have enough confidence to get on the plane."
As teams look to expand access to key systems and data, the ticket queue is pretty busy right now. But Maxwell is quick to share that the alternative — ungoverned AI building — would be a much greater challenge. "From a vibe-coding perspective, 3B makes our lives a lot easier,” he says. “People aren't trying to support a dozen different ways of deploying apps. It's all in one environment, where we can track who's building what, and what it touches."
3B has enabled people to do so many cool things. I think we might end up deprecating one of our core enterprise platforms this year, for example. 3B has been a huge productivity and collaboration boost for the entire organization.
Brandon Maxwell, Head of IT Operations & Information Security
AI agents play an increasingly large role in how teams at Tines work, with personal agents changing how people interact with existing tools and systems. Andrew on the security team, for example, built an agent that lets his team operate with the threat hunting capability of a team 10x their size. But deterministic builds provide a stable foundation for how work now gets done.
Up next: more building, more ambitious builds
Today, it's rare for any conversation at Tines to take place without a mention of something built with 3B. Teams are building at speed and volume — there are now 10,000 monthly runs of 3B builds for every Tines employee. Internal hackathons help employees build their confidence, while team-specific office hours provide a forum for troubleshooting and sharing inspiration.
I don't think there's any way anybody at Tines would want to go back to the way we did things before. Everyone is moving faster. Expectations have changed. The power people have now to own their own jobs has changed entirely, and nobody wants to give that up.
Martin Moroney, Head of AI Transformation
For Peter Coleman, and for a growing number of people across Tines, the old way is already hard to remember. His billing platform cost around $1,000 to build and can run indefinitely without ever calling an AI model again. It also removed the need for a vendor-supplied tool, which his team had been quoted $100,000 annually for. But for Coleman, the savings aren't even the best part. He says, "If I could have any billing platform on the market for free, I'd still choose this one, every time.”