Our manifesto: seven ideas for engineering that leads the way

We need to build more, build better and leave less room for error. More housing, more data centres, more infrastructure, on tighter schedules and with budgets that don’t forgive a bad early decision. And yet MEP and structural engineering still works the way it did twenty years ago: it comes in late, works by hand and reaches the end of the project with the margin already spent.

For ten years we’ve engineered MEP and structural systems for demanding developers and public institutions, and for the last two we’ve used AI on every project. What we’ve learned fits into seven ideas. They guide what we do, and what we don’t.

(01) Engineering can’t be the bottleneck.

On too many projects, engineering is what keeps everyone else waiting. Every architectural revision opens up weeks of recalculation, and every week of recalculation is a decision made without data. We want it the other way round: engineering ahead of the project, turning into the developer’s advantage.

(02) What matters is decided early.

The HVAC system, the MEP routing and the structural scheme lock in much of a building’s cost and energy use for decades. Those decisions are made at feasibility and concept stage, while every euro can still move. That’s where we want to be: comparing far more options, with cost, energy use and schedule on the table, rather than arriving once the design is closed.

(03) AI does the work. We sign it.

Apogea OS, our AI system, reads models and documents, sizes, routes and tests options. It does in hours the repetitive work that used to take weeks. But it doesn’t decide for us, and it doesn’t answer to you: every deliverable is reviewed and signed by a named Apogea engineer. Being AI-native isn’t about handing over judgment; it’s about freeing it up for what matters.

Judgment stays human. The repetitive work doesn’t.

(04) A change isn’t a restart.

Projects change: the brief, the end client, the regulations, the budget. That isn’t a failure; it’s reality. What makes no sense is for every change to cost as much as the first design. When the model changes, our system recalculates airflows, loads and circuits and gets everything ready for review. The change is absorbed; the project moves on.

(05) A building isn’t finished at handover.

Most of a building’s life begins when the lights come on. That’s why we work with a single coordinated MEP and structural model from day one, and that model doesn’t die on site: it becomes the foundation for maintaining, refurbishing and valuing the asset.

(06) Building data belongs to the developer.

Your building will outlive any software licence. That’s why we build on open standards and formats: your models and data are yours, they open in any tool, and you can take them with you whenever you like. Not locked to a vendor, and not locked to us.

(07) We answer to whoever carries the risk.

For years we worked as the architect’s subcontractor. We still work hand in hand with architects: they design, we make it work. The difference is that we now answer to those who develop and operate the building, with fixed scope, fixed price and payment for outcomes, not hours.

That’s what we mean by designing a common future: engineering that thinks about the whole building and its whole life, uses the best technology available without hiding behind it, and shares both information and responsibility with the developer.

If these ideas resonate with you, let’s talk.