The short version

Jack Dorsey’s latest idea, an “AI‑native company,” doesn’t mention rebar, RFIs or punch lists. But the story he’s selling should still make construction executives and project managers sit up.

At its core, the pitch is simple: instead of bolting AI tools onto old workflows, design the business so that AI is the default way work gets done. Data is captured for machines first. Decisions are simulated before they’re made. Automation isn’t a sidecar; it’s the engine.

For construction, which still runs on spreadsheets, siloed systems and gut feel, that’s a provocative contrast. Dorsey’s narrative is aimed at the tech crowd, but it sketches a future where AI in construction isn’t just a plug‑in for estimating software—it’s the operating system for the entire project.

An “AI‑native company” is a thought experiment that asks: what if you built the whole business around AI, instead of treating AI like an after‑hours add‑on?

Why this matters on real projects

Dorsey’s story lands in a world where construction technology is stuck between two eras. On one side: paper plans, phone calls, and the superintendent who “just knows” when a schedule is slipping. On the other: AI tools that can already read drawings, flag clashes, and predict risk—but are often used as pilots on one job, then quietly shelved.

The gap is not about algorithms; it’s about how companies are built.

An AI‑native mindset says: assume machines will:

• Ingest every drawing, spec, and RFI automatically.
• Run thousands of schedule and cost scenarios before a bid is submitted.
• Track field progress in near‑real time from photos, scans and sensor data.
• Surface early warnings on safety, delays and overruns without being asked.

Humans still make the calls. But the default is that automation does the first pass on everything that is repetitive, pattern‑based, or data‑heavy.

In practice, that could change familiar project moments:

• **Preconstruction**: Instead of a week of manual takeoff, an AI system chews through models and drawings in hours, then proposes multiple design‑to‑budget options. Estimators become editors and negotiators, not calculators.
• **Site coordination**: Daily photos and drone flights feed an AI engine that compares reality to the model. Deviations—out‑of‑tolerance embeds, missing sleeves—are flagged before they become rework.
• **Safety and logistics**: Cameras and wearables feed risk models. The system nudges the team when congestion, weather, or work sequencing make an incident more likely.

Dorsey’s storytelling doesn’t promise any of this to builders directly. But his "AI‑native" frame is a useful mirror: if a brand‑new contractor started tomorrow, with no legacy software, how much of their operation would they design around AI from day one?

That question is uncomfortable precisely because most firms are doing the opposite: layering AI tools on top of legacy ERPs, old habits, and fragmented data. It’s like hanging smart sensors on a tower crane powered by a diesel engine from 1978.

The tension in Dorsey’s narrative—between what’s possible if you start fresh and what’s messy inside existing companies—is exactly the tension construction leaders now face. The opportunity is enormous, but only if AI in construction is treated as infrastructure, not as an app.

What to watch next

• **AI‑first project pilots**: Owners and GCs experimenting with projects where schedule, cost control and quality workflows are designed around AI and automation from day one, not added mid‑stream.
• **Data foundations**: Quiet but critical investments in cleaning up drawings, models, and field data so AI tools have something trustworthy to learn from.
• **New roles on site**: The rise of AI coordinators or "digital supers" who sit between field crews and construction technology platforms.
• **Contract language**: Emerging clauses that address AI‑generated insights, responsibility for automated decisions, and how data from jobs can be reused.
• **Upstart competitors**: New firms that position themselves as AI‑native contractors or subs, using lean teams plus heavy automation to undercut traditional players.

Field note from the editor

I’ve walked jobs where the most advanced tech on site was a cracked iPad and a printer in a shipping container. I’ve also seen small teams use a handful of AI‑powered tools to run circles around much larger competitors.

Dorsey’s “AI‑native company” is a story aimed at Silicon Valley, but it’s a useful provocation for construction. The real question isn’t whether you’ll use AI. It’s whether you’ll keep treating it like a gadget—or start reshaping your business so that, one day, it feels like the way you always built.