Everyone in the world of business strategy knows The Innovator's Dilemma. Clayton Christensen's seminal work describes how great companies fail. In summary, it's not because they do things wrong, but because they continue to do everything right. They listen to their customers, invest in their most profitable products, and optimize their operations. And in doing so, they leave themselves vulnerable to smaller, scrappier disruptors who attack from below.
Fewer people talk about Christensen's follow-up, The Innovator's Solution. The solution is that, to survive disruption, a company must approximate the conditions of those who would disrupt it. It requires creating independent, small, unaligned teams that function like a startup within the organization, teams that are free to ignore the rules that make the main business successful. In late 2025, we set out to do exactly that at Tines.
On paper, this move looks irrational. Tines is in a strong market position. Our customers are happy and our product team is shipping fast. There were probably folks at Tines who didn't understand why we (Stephen, Head of Product, and Eoghan, Lead Product Designer) would step away from those roles to work in isolation on a yet-to-be-determined project.
Why take two people so deeply embedded in our product and cast them into the unknown?
It may have seemed like an impulsive decision, but it really wasn't. For a while, the leadership team had felt an itch — a sense that teams could use an AI-native tool to build apps, agents, and automations securely. There had even been a few earlier attempts to build something like this, but the technology wasn’t ready. All that changed in 2025.
Everyone’s a builder now
In 2018, when Tines was founded, building an automation meant either writing code yourself or relying on a tool that abstracted the code away for you. By late 2025, that barrier had almost disappeared.
Tools like Claude Code and Codex allowed non-engineering teams to describe what they wanted and have an AI model write the code to build it. In response, many IT and security teams put up a different kind of barrier: locking down access to sensitive data and systems before employees could connect to them and inadvertently put the organization at risk. But locks don't always hold — employees who hit a wall often just paste their own usernames and passwords straight into a vibe-coded script to get around it.
For IT and security teams, this is a big problem, and a fast-moving one. More people building with AI means more risk to manage and less visibility over that risk. We call this phenomenon wild code, because of how it takes root, spreads, and threatens to introduce vulnerabilities, blow through AI budgets, and overwhelm teams responsible for keeping enterprise systems safe.
But locking things down also comes with risk. Too much friction, and you block the kind of AI experimentation that many organizations are now mandating. IT and security teams find themselves stuck between two failure modes — enable too freely and expose the business, or restrict too tightly and kill the momentum. This tension is central to Tines 3B.
At the same time technology was creating this problem, it was also handing us the means to solve it.
AI was helping our own product team move faster than ever, and it felt like the right moment to try again. So, in 2025, we decided we would step away from our regular roles to work on a project with no roadmap, no deadline, and no defined outcome.
A third baby
Every project needs a forcing point. Ours was more coincidence than plan: we were both due to welcome new babies to our families and take paternity leave around the same time. We wrapped up our existing projects, spent some time with the new arrivals, and came back to work in October 2025 ready for a fresh start. We named the project 3B for Third Baby. At the time, this new piece of software felt like exactly that: a needy newborn, demanding lots of energy and attention.
We opened a Slack channel on October 1st. It became, for six months, our only shared record of the entire project. There were no strategy docs, not even a to-do list. We had the luxury of shedding existing processes and recurring meetings. Thanks to the exceptional team who stepped in to continue working on Tines — now called Tines Stories — we were able to focus entirely on building this new product.
We set ourselves some deliberately absurd goals in our first week. The first: build something 100 times simpler to use and 100 times more scalable than Tines Stories, which was already one of the most powerful workflow platforms on the market.
The second: have a working prototype we could show a customer by the end of 2025. After two weeks of working together, these goals didn’t seem so absurd anymore.
An unfair advantage
We didn’t have a plan, but we had something better — eight years of experience helping customers at top companies power their most important workflows. In our case, "starting from scratch" meant starting with thousands of customer conversations already behind us, and a clear sense of what makes a product customers genuinely love.
It also meant we didn’t struggle to find our first users. We had Tines Stories customers willing to take a look at early versions of Tines 3B, and power users inside the company ready to poke holes in our assumptions as we built. Getting that kind of validation and feedback before a product is finished is a luxury most new startups can only dream about.
Monitoring in Tines 3B
Building in Tines 3B
Collapsing the stack
We went into Project 3B knowing that a lot of what we'd built since 2018 — the abstractions, the visual metaphors designed to make automation approachable without code — would no longer be necessary.
We started with what felt like an extreme idea — collapsing Tines Stories' eight action types down to three. We ended up somewhere more extreme still: just code.
That same logic ended up collapsing something else: the line between our own roles. In December 2025, a step change in coding models meant AI coding tools no longer needed constant engineering fixes.
Eoghan, a designer by training, tried it for the first time on something trivial: resizing a header. The change went live, shipped without an engineer in the loop. That small fix triggered weeks of obsessive building, and suddenly Eoghan could ship a fully realized front end himself — sound design, micro-interactions, the polish that usually dies in a Figma file. Almost overnight, he became a back-end builder too, moving fluidly between product design and production code in a way that simply wasn't possible a year earlier.
Freed from building every pixel himself, Stephen went deep on the part that keeps us up at night as a security-focused company: making sure AI writing and running code on someone's behalf does so safely, with the right boundaries and governance in place. This governance layer is what sets Tines 3B apart from the wave of AI coding tools appearing elsewhere in the market.
Opening the floodgates
By March 2026, we were ready to show the company what we'd been working on. Tines 3B was demoed on stage at our company kickoff in Galway, Ireland, in front of about 400 employees. We tried the same use case we'd demoed in Tines Stories at a company kickoff two years earlier — only this time, we built it in a fraction of the time.
We didn't yet have an answer to the obvious question everyone was asking: what is this? We laid out the honest possibilities — Tines 3B could live alongside Tines Stories, fail outright, or fold back into the core platform. We genuinely didn't know.
But that uncertainty didn't last. After CKO, our sales team immediately started asking to show Tines 3B to customers. The response was overwhelmingly positive.
Customers weren't just interested in Tines 3B, they were already asking when they could get their hands on it.
This is also when the project stopped being a two-person effort, as we needed enterprise fundamentals like single sign-on and role-based access control.
The team grew, one person at a time: Julia Grabos, Joe Donovan and Shayon Mukherjee in engineering, Colin O'Brien from our security team, and Rosie Ellison as product manager. None of this came at Tines Stories' expense — the pace of hiring at Tines meant we could build out the Tines 3B team without pulling investment away from the product our customers already depend on every day.
Built for trust
We believe our customers are experts in their own work, and that work is too complex and too specific to force into someone else's assumptions. That's why Tines 3B is built around primitives — code, connectors to the systems you already use without being locked into any one vendor, and a simple way to store data — rather than a pile of narrow features stacked on top of them. This way, if you want to change how a workflow behaves, you're not waiting on us to ship a feature for it.
So what did we build? A product that:
Prioritizes security in its architecture, feature set, and outputs
Runs reliably at scale
Is intuitive and easy to use, regardless of technical ability
Enables collaboration and sparks creativity
Amplifies user impact with AI
That philosophy shaped what we didn't build, too. We didn't want to lock customers into one AI provider, add a dumbed-down "easy" mode, or allow feature sprawl to erode the power underneath. What makes Tines 3B different from a consumer AI-building tool isn't the building — plenty of tools can do that now. It's that an organization can trust it.
Employees are already building like software engineers, with or without their company's permission. The real question was never whether to enable that. It was how to make it safe for an enterprise, instead of quietly exposing the business to risk.
That trust, earned through governance and security built in from day one, is what makes Tines 3B reliable at real scale today — not just for a side project, but for the workflows a business actually depends on.
The next generation of Tines
We started this project not knowing where it would end. Today, the answer is clear: Tines 3B is the next generation of Tines.
Tines Stories continues in parallel, trusted by some of the best enterprises in the world. Tines 3B builds on that same trust. It offers an AI-native environment for every team's agents, apps, and automations to run securely at scale.
From the beginning and throughout Project 3B, our mission at Tines has never changed. We're still here to power the world's most important workflows. We just found a new way to do it.