The $13 Trillion Industry
Still Running on Copy-Paste

I've watched a senior architect spend 45 minutes reformatting a BOQ spreadsheet that should have been auto-generated from the model she'd already built. Forty-five minutes. Not designing. Not thinking about load paths or daylighting or the client's brief. Just copying cells from one tool into another, double-checking unit conversions, and praying nothing got lost in translation.

She's not bad at her job. She's one of the sharpest designers I know. And that's exactly the problem.

The Dirty Secret Nobody Talks About at AEC Conferences

The architecture, engineering, and construction industry is worth roughly $13 trillion globally. It employs hundreds of millions of people. It shapes every city skyline, every hospital corridor, every bridge you drive across.

And yet — and I say this as someone who's spent years in this space — the average schematic design workflow is held together with duct tape.

Here's what a typical early-stage design process actually looks like at most firms:

You get a brief. Maybe it's a PDF. Maybe it's a rambling email with an attached site plan from 2019. You open one tool to sketch out the massing. Another tool for the floor plans. A third tool for renders — because the client won't approve anything without a pretty picture. Then you hop into a spreadsheet for area calculations. Another spreadsheet for cost estimation. Maybe you export an IFC and it's missing half the metadata because the export settings were wrong and nobody caught it until the QS called three days later asking where the slab data went.

I Counted the Tools Once. I Stopped at Eleven.

For a single residential project — nothing fancy, a mid-rise apartment scheme — I mapped out every piece of software the team touched between receiving the brief and submitting the Stage 2 package:

Eleven tools. And the data flowed between them via the most ancient integration protocol known to mankind: a human being manually retyping things.

Nobody designed this workflow. It just… accumulated. Every firm has its own version of it, and every architect silently accepts it as the cost of doing business.

Where the Hours Actually Go

McKinsey published research showing that large construction projects typically run 80% over budget and 20 months behind schedule. The Economist has covered the productivity gap in construction multiple times — it's one of the least digitised major industries on the planet.

But let's forget the macro numbers for a second and talk about what happens inside a 15-person architecture studio on a Tuesday afternoon.

The cost estimation loop. An architect finishes a massing study. The QS needs quantities. So the architect exports. The QS imports — or more likely, rebuilds the quantities manually because the export was missing context. They plug numbers into a rate sheet. Two days later, the estimate comes back. The architect changes the massing. The QS starts over.

The render bottleneck. A client meeting is Thursday. The design partner wants three concept options visualised. Each one needs to be modelled, textured, lit, and rendered. That's a minimum of 6–8 hours per option if you're using a traditional rendering pipeline. So the team works late on Wednesday. Again.

The IFC black hole. The structural engineer needs an IFC. The architect exports one from Revit. The engineer opens it in Tekla. Half the walls are classified wrong. The property sets are empty. Someone has to go back and fix the mappings. This happens on almost every project, and it's been happening for fifteen years.

So What Would It Look Like If You Actually Fixed This?

I've been thinking about this question for a long time. And recently I came across something that made me stop and pay attention.

It's called BuildFlow, and the simplest way to describe it is: a workflow engine built specifically for AEC.

Not a new CAD tool. Not another rendering app. Not a project management platform pretending to understand architecture. It's a pipeline builder — you connect nodes together to create automated design workflows, and the AI handles the translation between stages.

Think of it like this: instead of opening seven different tools and manually transferring data between them, you build a pipeline once. Text brief goes in one end. 3D massing, renders, floor plans, IFC files, cost estimates come out the other end. The data flows through the pipeline without you having to touch it.

I'll admit I was skeptical at first. I've seen plenty of "AI for architecture" pitches that amount to a chatbot generating floor plans that violate basic building regulations. BuildFlow is different because it doesn't try to replace the architect's judgment — it eliminates the busywork between decisions.

What This Actually Looks Like in Practice

Let me walk through a few of their workflow templates, because the specifics matter.

Why This Matters Beyond Saving Time

I know what some people are thinking. "Sounds like a shortcut. Where's the design quality?"

Fair question. Here's my take.

BuildFlow isn't generating final construction documents. It's operating in the schematic design phase — RIBA Stages 0 through 2, if you're in the UK, or SD/DD if you're in the US. This is the phase where you're exploring options, testing feasibility, communicating ideas to clients, and iterating quickly.

When it takes a full day to produce one concept option, you only show the client two. When it takes 90 seconds, you show them ten. That's not lower quality — it's radically better design exploration.

And the data integrity issue is real. Every time a human manually transfers a number from one tool to another, there's a chance of error. BuildFlow's pipeline approach means the data flows through nodes programmatically. The massing dimensions that inform the cost estimate are the same dimensions the system used to generate the 3D model. No copy-paste. No transcription errors. No "wait, did we use gross or net area?"

The Bigger Picture

The AEC industry has a productivity problem that's been documented for decades. Every other major industry has undergone significant digital transformation. Manufacturing has automated assembly lines. Finance has algorithmic trading. Healthcare has electronic records and diagnostic AI. Construction? We've got BIM, which is powerful but only addresses part of the workflow. The gaps between tools remain.

What BuildFlow is building — a node-based, AI-powered workflow engine for the entire AEC pipeline — feels like the missing layer. Not replacing Revit or Rhino or ArchiCAD. Complementing them. Automating the boring connective tasks so that the people using those tools can focus on what they were actually trained to do.

They currently have 13 workflow templates covering concept design, visualisation, BIM export, cost estimation, and 3D modelling. They've got 30+ AI nodes, and the community can request and vote on new workflows. The pricing starts at free — three executions per month — and goes up to ₹1,999/month for the pro tier with 100 executions and priority support.

Is it perfect? No. It's in beta. Some nodes are still in preview. The community hub is still growing. But the approach is right. The architecture of the platform — pun intended — is sound.

What I'd Tell My Firm

If I were running a practice right now, here's what I'd do:

I'd take one upcoming project — something in early schematic design — and run it through BuildFlow's "Text → Floor Plan + Render" branching workflow alongside the traditional process. Same brief, same timeline. Then I'd compare: how many concept options did each approach produce? How long did each take? How accurate were the area calculations?

I suspect the results would be uncomfortable. Not because BuildFlow would produce better designs — that's still the architect's job — but because it would expose how much time the traditional approach wastes on tasks that have nothing to do with design.