How Much Should a Structural Engineer Cost?

At an industry event we attended recently, one of our design partners raised a worry that has stuck with us since. AI tools are starting to land in structural engineering workflows, and the early productivity gains are real. The concern is what happens next.

If a piece of work that used to take a week now takes three days, the engineer does not capture that gain. The client does. What was once a reasonable timeline for a structural package becomes the new minimum, and the next negotiation starts from there. The engineer is doing the same work to a higher standard in less time, and the value of that work, as priced by the market, has fallen.

The worry is that this is a race to the bottom. While the engineer's intrinsic value as a designer of safe and habitable buildings has not changed, the price the market is willing to pay for that value, measured in hours, has. In a discipline that already runs on thin margins, a steady compression of fees is an existential risk.

This is not a new problem. Economists have been arguing about whether labour-saving technology raises or lowers the value of skilled work since the nineteenth century, and the question has yet to be settled in any general way. 

What is actually happening in the discipline is that the work has expanded faster than the pricing model has been able to absorb. The Building Safety Act has loaded an enormous amount of additional rigour onto RIBA Stage 4 work. The SER scheme in Scotland imposes audit and record-keeping obligations that did not exist twenty years ago. Engineers are doing more, with more accountability, under more scrutiny, for fees that were agreed against a scope that no longer matches the job.

If AI saves engineers time, it does not relieve any of this. It absorbs into a workload that was already growing. The hours come down. The risk does not.

Structural engineering has been priced on time for as long as anyone can remember. Hourly rates, day rates, fee scales tied to a percentage of construction value. All of these are proxies for the time the engineer is expected to spend on a project.

Time was a workable proxy when the work was reasonably stable. A structural package of a certain scope took a known number of engineer-weeks. The price reflected the time. It was rough, but it worked.

It doesn’t work quite as well now. Two engineers running the same project on the same scope can spend wildly different amounts of time on it, depending on how many design changes come in, how much downstream propagation is required, and how rigorous the audit trail has to be. Time is no longer a clean measure of effort. And effort is no longer a clean measure of value.

What an engineer actually delivers, beyond the drawings and the calculations, is something closer to insurance. They put their name on a structure that the public will use. They sign off on compliance with codes that carry personal criminal liability. They absorb design changes from every other discipline and make the building work. The thing they are selling is, in part, the protection of downside: the building does not fall down, the firm does not get fined, the client does not lose a year to a Gateway rejection.

That is a different thing from time. It is harder to measure, but it is closer to what the work is actually for. A junior engineer's hour and a chartered engineer's hour are priced differently for a reason, but even that distinction is a coarse approximation of value. The work that goes into a Gateway 2 submission, the precedent search that surfaces the right detail for a listed refurb, the change-impact analysis that catches a problem before it reaches a contractor: none of these are well-described by their duration.

In conversations with our design partners, two threads keep coming up. The first is margin erosion: firms are paid the same as they were five years ago for work that has materially expanded in scope and rigour. The second is risk: the personal and professional consequences of a missed change, a gap in the audit trail, or a non-compliant signature have become severe enough that they need to be priced explicitly rather than absorbed silently.

A value-based pricing model would price both of these directly. The engineer's fee would reflect the downside they are protecting against (Gateway 2 enforcement, professional indemnity exposure, reputational damage on a flagship project) and the upside they are creating (faster gateway approvals, fewer rework cycles, a defensible audit trail when the client comes back in five years). When the work expands faster than the clock, the clock stops being the right unit.

Renegotiating the fee model on a structural commission is not something most firms can do unilaterally. The contract is signed, the scope is defined, the rate is agreed. Telling a developer halfway through a project that you are switching to value-based pricing is a quick way to lose the next project.

Closer to home, AECOM is doing this in real time. In November 2025, the firm acquired Norwegian AI start-up Consigli AS for $390 million and built its tools, which Consigli claims can cut engineering time by up to 90%, into AECOM's design workflow. The interesting move was not the acquisition itself. It was what AECOM did with its contracts. On their Q2 2026 earnings call, the firm disclosed two project wins worth nearly $1 billion structured around commercial mechanisms that let AECOM share in the value created by AI-driven efficiencies, rather than passing the entire benefit to the client through compressed billable hours. AECOM called it "gain share." In practical terms: when AI made the work faster, AECOM's contracts let the firm keep some of the productivity gain rather than absorbing it as fee compression.

While AECOM is a $16 billion global firm and most structural engineering practices are not, the principle is the same at any size. The tractable place to start is the design change.

Every change that comes in during a project carries a real cost. The model has to be rerun. The drawings have to be updated. The calculation pack has to be revised. Downstream documents have to be regenerated. Coordination with the architect, the MEP team, and the contractor has to be redone. Most of this work is currently absorbed silently by the structural team, with no record of what each change cost and no specific charge against it.

If a firm can quantify what a requested change actually costs, including the engineer-hours it consumes, the propagation it triggers, and the audit trail it forces, and put that number in front of the client at the moment the change is requested, the conversation changes. The client is no longer asking for a small tweak. They are looking at the cost of additional scope. Some changes will still be absorbed, but many of them should not be.

This is not a new fee model. It is the practice of putting a price on a category of work that the industry currently treats as free. It ties the cost to a specific, evidenced piece of work rather than to an abstract argument about what structural engineering is worth.

There is a reason the shift to value-based pricing has been slow in structural engineering, and it is not the engineers. We spoke recently to a developer who put it bluntly. Their costs are tied to the repayment value of the loan they have taken out to deliver a project. Time is the variable that defines their financial exposure. Every week of delay costs them measurable money, and every week saved is measurable margin.

That is a perfectly rational position, but it means the developer's incentive structure pulls the entire design chain back towards time-based pricing. The architect quotes against time. The structural engineer quotes against time. The MEP consultant quotes against time. Everyone is being measured against a clock that was set by the loan agreement, not by the work.

The shift towards value-based pricing cannot happen one node at a time. It requires the developer, the principal designer, and the structural engineer to agree on what the work is actually worth, what risks are being absorbed, and who is bearing them. While it is a harder conversation than asking for a higher day rate, it is the only conversation that ends in a sustainable answer.

The near-term version of this conversation does not require the whole chain to move at once. If the structural engineer can show, change by change, what each piece of scope creep is costing them, the developer's time-based framing starts to break down on its own terms. The developer can still choose to push the change through. They are just doing so with a number attached, rather than as a free option.

The interesting thing about AI in this context is that it is not really the threat. The threat is the pricing model. AI exposes the inadequacy of time-based pricing because it accelerates the disconnect between hours spent and value delivered, but the disconnect was already there.

If an engineer uses AI to absorb a Gateway 2 documentation burden that would have taken a junior engineer six weeks, the value of that absorption has not changed. The building is no less safe. The audit trail is no less defensible. The firm has just done the work with less consumption of billable hours. Under a time-based model, the engineer earns less. Under a value-based model, the engineer benefits from every productivity gain rather than being penalised for it.

Tooling does not solve the pricing problem on its own, but it does make the value visible. When a firm can show, with evidence, how many changes it absorbed, and how long each one took, the conversation with the client stops being conducted on assertion. It is conducted on evidence. And evidence is what makes value-based pricing possible, starting with the changes and working outwards from there.

The question of how much a structural engineer should cost is not one that gets answered by a blog post. But it is one the industry will have to answer in the next few years, because the alternative is firms doing more for less, with more risk attached, until the economics give out.

If your firm is thinking about any of this, we would love to hear how you are approaching it. Get in touch at [email protected], or reach out to any of us directly. Part II coming next week.