A client arrives with approved drawings, a settled budget and one expectation: that somebody confirms. A serious engineering office starts by answering something else — what are your input data? Without a ground investigation, without defined loads, without a stated use, there is nothing to confirm. That polite refusal is the first professional act of the trade.

The technical third party on a project

On a construction project, three roles are distinct and should not be conflated. The designer defines what is to be achieved. The contractor decides how to achieve it. The engineering design office establishes what that requires technically, and sets it down in documents that engage its liability.

Its clients are not always the same either: it works for architects, for construction companies, for public or private clients. It sometimes comes in very early, on a sketch, to say whether an intention is feasible before the drawings are fixed; sometimes later, from settled drawings.

That third-party position has a useful consequence: the engineering office has no interest in the project looking simpler than it is, nor in it looking dearer. Its value rests precisely on having nothing to sell beyond its analysis.

Nothing starts without input data

A technical study is not an opinion: it is reasoning applied to data. If the data are missing, the reasoning is worth nothing, however competent the person conducting it.

The input data are the building's real use and the loads that follow from it, its geometry, the nature of the ground, the climatic conditions and site hazards, the materials contemplated, the execution constraints, and the applicable regulatory framework. Each of these is documented; none is assumed.

This is also the stage at which the quality of the dialogue with the designer is decided. A grid revised in time, a load path anticipated, a plant room moved two metres cost one conversation. The same adjustments after the drawings are approved cost a complete redo.

The ground decides more than the drawing

This is where an engineering office most clearly distinguishes itself from an obliging supplier.

The ground is not one parameter among others: it governs the type of foundation, and therefore a large share of the cost and most of the risk. Two neighbouring plots can behave very differently. Geotechnical investigation is the only way to know anything about it, and it belongs to a laboratory or a specialist whose trade that is — not to a visual impression on site.

An engineering office can coordinate that work, use its results and draw the consequences. What it does not do is size foundations from unverified assumptions. There is no depth, section or bearing capacity that applies by default: those values follow from a ground, a structure, loads and standards, and a discovery article has no business proposing any.

Sizing means arbitrating

Structural analysis is not a single operation producing one answer. It is a series of trade-offs between safety, feasibility, cost, time and ease of construction.

The same building can often be resolved in several ways, with different consequences for the quantity of material, the complexity of the formwork, the time on site and what the contractor can actually do. The engineer's role is to choose a defensible solution and to say why.

The deliverable that embodies that work is the calculation note, together with the corresponding drawings. That document is dated, referenced, based on written assumptions and signed. That is what distinguishes it from a verbal opinion: it binds whoever produced it, and it can still be relied on years later.

Deliverables a contractor can actually build from

A study that is correct but unusable protects nobody.

The quality of a deliverable is measured on site: are the drawings precise enough for the tradesperson to work without phoning? Are the references consistent between documents? Do the schedules match what is drawn? Are the openings for services carried through everywhere they need to be?

An incomplete set does not produce one visible error: it produces local interpretations, taken by different people at different times, which only contradict one another once the concrete is poured.

Digital models and clash detection

Three-dimensional modelling and digital models changed one specific thing: they let packages meet each other before the site does, where they used to discover each other in place.

A beam crossing a ventilation duct, a downpipe landing in an opening, a suspended ceiling incompatible with the available height: these clashes already existed, but they were found by running into them. Detecting them upfront removes rework and weeks of delay.

No particular software is a requirement of the trade, however. What matters is effective coordination between packages and the ability to pass on the result in a format the other parties can open and use.

Standards and codes: a framework that changes at every border

The requirements applicable to a structure — stability, fire safety, escape, accessibility, technical installations, resistance to local hazards — depend on the country, the building's use, its size and its location. The reference standards used also differ from State to State, and an engineering office works within the regulatory framework of the place where the structure is being built.

The United Nations Office for Disaster Risk Reduction points out that resilience must be built into every stage of an asset's lifecycle, encourages States to strengthen their national and local regulatory frameworks, and has developed with more than a hundred countries a set of principles for resilient infrastructure. It also notes that building resilience into infrastructure adds only about 3 per cent to total investment costs, an amount easily offset by the losses avoided.

How the professionals fit together is decisive here. The International Union of Architects addresses scope of practice explicitly in its Accord adopted in Beijing in 1999: knowing what belongs to design, what belongs to analysis and what belongs to construction avoids the grey areas in which failures take root.

Checking during construction

A study that is never verified on the ground protects only on paper.

Visits at key stages serve to confirm that execution matches the documents: positions, visible sections, constructional arrangements, respect for the assumptions adopted. They give rise to a written report, which is worth more than a remark made verbally to a foreman.

These visits take place on a live site, with everything that implies. Convention No. 167 of the International Labour Organization on safety and health in construction, adopted on 20 June 1988 and in force since 11 January 1991, covers all construction activities from the preparation of the site to the completion of the project: safety therefore concerns those who come to inspect as well as those who build.

Performance, materials and service life

Engineering offices are drawn in ever earlier on questions that historically did not belong to calculation: a building's expected consumption, its thermal behaviour, the choice of materials with respect to installation and upkeep, water management, the service life of the works.

The United Nations Environment Programme, which hosts the secretariat of the Global Alliance for Buildings and Construction, states that buildings account for around 37 per cent of energy and process-related CO₂ emissions and over 34 per cent of energy demand, and runs work on circularity approaches in the built environment.

Here too, no solution holds everywhere: what is relevant depends on the local climate, the materials available, the skills present and the applicable regulations.

What artificial intelligence does not sign

Assistive tools are useful for organising data, comparing variants, preparing a note or sorting a technical file. They save time on what is repetitive.

They validate no structure, certify no calculation, replace no ground investigation, guarantee no compliance and issue no permission. Above all they sign nothing: responsibility stays with the professional who puts their name on the document.

What an engineering office commits to, and what it does not guarantee

An engineering office does not guarantee that a structure will never develop a defect, that a contractor will build in accordance with its drawings, that a budget will hold, that a permission will be granted, or that ground will behave exactly as the boreholes suggested. It does not build and does not supervise the site continuously.

What it does commit to is precise: written assumptions, a study based on real data, dated and signed documents, usable deliverables, coordination genuinely taken on with the other parties, and inspection reports where the appointment provides for them.

That is also what makes an office legible to those who commission it: a stable professional identity, fields of work described without exaggeration, appointments priced before commitment, deliverables presented for what they are, and archiving that allows an assumption to be retrieved ten years later.

Finally, this article describes a trade; it contains no sizing recommendation and replaces no study. The requirements mentioned vary from country to country, and any real structure is dealt with on data, with the qualified professionals where it is being built.