Why I left big tech to build something that actually fits your workflow
Are you prepared for what’s next in AECO?
I spent twenty years at large technology companies. That experience taught me a lot about building products at scale, as well as the pressures that can make leaving a platform harder than it should be.
It starts with a proprietary data format. Then an acquisition that removes an integration. Then a pricing change that makes switching painful. Then an AI layer that's only useful if you let the vendor train on your data. At each step, the friction of leaving increases. At each step, you have a little less control over your own information.
That's not a partnership. It's a trap with good UX.
When I looked at Revizto from the outside before I joined, what I saw was different. A company that had spent fourteen years building a platform that worked across formats, connected to other tools, and kept customer data exactly where it belonged — with the customer. I joined because I wanted to build something that earns its place in your workflow every single day. Not because switching is too hard.
What 'open' actually looks like in practice
One of the most significant announcements I made at Made Right 2026 was the launch of our read-only API for object properties, with full model access following within 24 hours.
This is the foundation for everything we're building toward. Your project data, your issue history, your model properties, your coordination records — all of it accessible to your tools, your workflows, your AI of choice.
The use cases are immediate and practical. Quantity takeoffs that pull live from the model. Procurement systems that receive quantities directly. Quality audits that run automatically against your model. The reason this matters is not just efficiency. It's control.
The AI question
We are not building a closed AI system that trains on your data. We are not locking you into a proprietary model. We are building the infrastructure for you to bring your own AI to your own data. You can read exactly how we approach data security and sovereignty — and what our certifications actually mean in practice.
The MCP integration we demonstrated at Made Right connects the Revizto Collaboration Hub to Claude, ChatGPT, Gemini, or whatever tool your organization has already approved. You ask questions in natural language. You get answers drawn from your live project data. You stay in control of everything.
That position doesn't change depending on which AI tool is popular this quarter. It's a principle — and you can read our full thoughts on it in Revizto's AI Manifesto.
FAQs
An open API in construction technology means project teams can extract their own data — including model properties, issue history, and coordination records — and connect it to external tools such as procurement systems, ERP platforms, quantity takeoff software, or AI tools of their choice. Without open API access, that data remains inside the vendor's platform, limiting how it can be used, analysed, or integrated across an organization's wider technology stack.
MCP (Model Context Protocol) integration allows construction software platforms to connect directly with large language model AI tools such as Claude, ChatGPT, and Gemini. Rather than building a proprietary AI layer trained on customer data, MCP integration lets project teams query their own live project data using natural language, through whatever AI tool their organization has already approved and deployed.
Construction teams should ask whether a vendor offers open API access to project data, supports integration with third-party tools without restrictions, and commits to not using customer data to train proprietary AI models. Vendors who describe themselves as open but restrict data portability, limit integrations to approved partners, or require data sharing as a condition of AI features are not offering genuine openness — they are managing the terms of lock-in.
ISO 19650 is an international standard for managing information over the whole life cycle of a built asset using building information modelling. Construction software that supports ISO 19650 naming conventions allows project teams to maintain consistent, compliant document and model management across disciplines and organizations, reducing the risk of coordination failures caused by inconsistent file naming and version control.
A closed AI system in construction technology is one where the vendor builds and controls the AI model, typically by training it on aggregated customer data. A bring-your-own-AI approach, by contrast, gives project teams the infrastructure to connect their own project data to whichever AI tools they have chosen — without the vendor accessing, retaining, or benefiting from that data. The distinction matters because it determines who controls the intelligence derived from your project history, and who owns the value it generates.
