In construction, precision is not a virtue — it is the baseline. The axis of a reinforced concrete column tolerates a few millimetres of deviation. The concrete mix is measured to the gram. The structural calculation admits no compromise. Anyone who spends decades in the trade learns a kind of precision few industries demand so consistently.
Then that same person walks back into the office and spends three days hunting for the e-mail attachment holding the performance certificate a subcontractor sent two weeks ago. The moment that precision leaves the site, it simply evaporates.
Not because nobody cares. Because until now, nobody handed them the tool for it.
01 — The diagnosisThe gap is not in the trade
A mid-sized general contractor runs eight to ten projects at once. Twenty to thirty subcontractors per project. Per subcontractor: a contract, amendments, performance certificates, deductions, retention, warranty obligations. The product of those numbers is already in the thousands — and all of it typically lives in a shared folder, seventeen spreadsheets and a few thousand e-mails.
The system works. It just happens to work inside one person’s head, and when that person goes on holiday, the company grinds half to a halt.
Construction is not slow. It has simply been given software that never spoke its language.
The market has offered a remedy for decades: general-purpose enterprise systems. They can do everything — from stock levels to payroll. There is exactly one thing they cannot do: know what a performance certificate is. What a reverse-charge VAT subcontractor chain is. What retention is, and when it falls due. Those concepts have to be built into them during an implementation project, over months, at consultancy hourly rates — only for half the customisation to break at the next version upgrade.
02 — The propositionA system that knows the trade on day one
ATMOS starts from a simple claim: a construction company needs a construction system. Not a generic frame with some industry vocabulary bolted on — but core concepts that are already there in the data model.
Here a tender is not an “opportunity”, it is a tender: requests for quotation go out, bids come back, they are comparable against identical technical content, the decision is recorded, and the winning bid becomes a contract in one click. The contract becomes a performance certificate. The certificate becomes an invoice. The chain never breaks, because it does not live in separate systems.
- “Where is that certificate now?”
- Every approval round, every signature, every document version timestamped. Who approved what, when and from where — one click.
- “Just forward it to the subcontractor.”
- Nothing to forward. On the Stargate partner portal the subcontractor submits the completion themselves, and sees their contract, their deductions and open requests for quotation.
- “Draw up a contract from yesterday’s agreement.”
- Generated from parameterised templates — civil law, public procurement, reverse-charge VAT — versioned and signable via magic link. Twenty minutes of paperwork becomes twenty seconds.
- “What is this project’s actual margin?”
- Not something you learn at quarter’s end. Live cash-flow view, project profitability, warranty and bank guarantee tracking, partner ratings — continuously.
No single feature here is revolutionary. Their alignment is: that they work from the same data, on the same surface, at the same moment. On an iPad on site, on a laptop in the office, on a phone in the car.
03 — The proposition, continuedThe collective memory
A company’s most valuable asset is not its plant and machinery. It is the last ten years of priced tenders, signed contracts, and the subcontractors who delivered — or did not. That knowledge exists; it has simply been out of reach. It lived either in someone’s head or in an archived folder.
The ATMOS AI assistant — a module you can switch on — has access to exactly that asset. It is not a chat window bolted onto a system after the fact: it sees the live state of projects, contracts, bids and minutes. You can ask it in plain language. Which subcontractor had the highest reliability index last year? Compare these two quotations. Which approvals are due tomorrow?
The effect compounds: the data of every closed project becomes a more accurate price on the next tender. The company gets smarter year after year — not by hiring more people, but by not forgetting.
04 — TrustThe part nobody talks about
A construction company’s data is more sensitive than it tends to realise: margins, subcontractor terms, client contracts. Which is why ATMOS data lives on EU servers, in a Frankfurt data centre, travelling under TLS encryption, at rest under 256-bit encryption, with row-level permissions and a full audit trail. Not a clause at the end of a contract — an architectural decision taken at the very beginning.
The same goes for pricing: no implementation fee, modular build, updates included in the monthly subscription. If you don’t use fleet management, you shouldn’t pay for it.
For centuries, construction has done the most precise thing a person can do: build something that stands. It is high time its administration behaved the same way.
