When AECOM announced it was paying 4 billion Norwegian krone — approximately $390 million — for a five-year-old Oslo AI startup called Consigli in November 2025, the first question the AEC industry asked was: why is a services firm buying a software company?
Engineering services giants like AECOM — which reported $13.8 billion in revenue in fiscal 2024 — are in the business of selling the time and expertise of their people. Software companies sell licenses to tools that help other people do the work faster. These are not the same business model. They do not typically merge.
The second question, which followed immediately, was harder: if AECOM has paid $390 million for technology that its founder claims can automate 90% of engineering workflows, what does that imply about the value of the people whose work it is automating?
The answer to both questions is the same. AECOM is not trying to replace its engineers. It is trying to make them impossible to compete with.
What Consigli Built
Consigli was founded in Oslo in 2020 by Janne Aas-Jakobsen, a structural engineer with a background in project delivery on complex infrastructure projects. The company describes its core product as “The Autonomous Engineer” — an AI agent designed to handle the computational, iterative work that occupies a disproportionate share of an engineering firm’s billable hours.
The platform covers a specific and high-value slice of engineering workflow:
- Space analysis and unit optimisation — evaluating how building configurations trade off against programme requirements, running permutations that would take a design team weeks to work through manually
- Automated MEP loadings — calculating mechanical, electrical, and plumbing loads from building models without manual takeoff, a task that is simultaneously critical and time-consuming on every commercial and industrial project
- Level 3 modelling — advancing design models from concept to coordinated geometry, the stage at which the majority of engineering labour on a building project is consumed
- Plant room optimisation — finding efficient spatial arrangements for mechanical plant rooms, one of the most constraint-dense coordination problems in building services engineering
- Tender document risk review — scanning contract documents for scope gaps, ambiguities, and risk clauses that could generate claims
Consigli claims its system can reduce engineering time by up to 90% on these workflows and cut material usage by roughly 20% through optimised design. Before the acquisition, the company had collaborated with AECOM on several projects — the deal, it appears, was partially an outgrowth of that working relationship.
Janne Aas-Jakobsen will join AECOM as head of AI engineering following the acquisition.
Why an Engineering Firm, Not a Software Company
The normal exit path for a construction AI startup is acquisition by a software incumbent — Autodesk, Trimble, Nemetschek, Procore. They have distribution, they have the platforms the tools need to integrate with, and they have experience managing software businesses. A services firm acquiring a software company is structurally unusual.
The logic AECOM is applying is different. What Consigli has is not just a product — it is a capability. An engineering firm that can deliver the computational work of ten structural engineers in the time it currently takes two, at a lower cost, without sacrificing accuracy, has a structural advantage in project delivery that no software license can replicate.
That advantage compounds in procurement. AECOM competes for engineering mandates on the basis of team quality, track record, and price. If it can deploy AI that reduces the engineering labour content of a bid by 40% while maintaining quality — or use that labour capacity to take on more scope — the competitive implications for firms that have not made equivalent investments are significant.
The Verdantix analyst commentary around the deal was pointed: this acquisition “signals a step change where services giants are beginning to acquire AI capabilities and place innovation at the core of their business strategy.” The implicit question it raises for Jacobs, WSP, Bechtel, Arup, and every other major engineering firm is whether they need a similar move to remain competitive — or whether the software they are already licensed to use will be sufficient.
The Open Question That Makes This Deal Complicated
The most significant unresolved question about the Consigli acquisition is commercial: will Consigli’s Autonomous Engineer remain available as a standalone product to other engineering firms, or will AECOM use it exclusively?
This question has no confirmed answer. AECOM has not stated its intentions publicly.
If Consigli becomes an AECOM-exclusive tool, then the $390 million is a bet on internal productivity and competitive differentiation — a proprietary AI engineering capability that competitors cannot purchase. The risk is that the investment does not generate returns if AECOM cannot fully deploy the technology across its project portfolio at scale.
If Consigli remains commercially available, the question becomes whether AECOM is comfortable selling a core competitive capability to the firms it competes against. Some acquisitions attempt to preserve commercial product lines while restricting the most sensitive capabilities to internal use. Whether that balance is achievable with a technology as foundational as Consigli’s engineering automation platform is unclear.
The industry is watching the answer closely. Engineering firms that currently license construction AI tools from software vendors are evaluating whether those vendors — Autodesk, Trimble, others — will develop equivalent capabilities, or whether Consigli’s acquisition represents the beginning of proprietary AI differentiation that makes software licensing a secondary strategy.
What $390 Million Buys at a Five-Year-Old Startup
Consigli was founded in 2020, making it five years old at the time of acquisition. The $390 million price tag implies a substantial multiple on whatever revenue the company was generating — the financial terms of Consigli’s prior funding rounds were not publicly disclosed, and its customer base as a commercial entity has not been detailed.
The valuation reflects two things that are not on a typical financial statement: the engineering talent that built the system, and the training data and model development that went into the Autonomous Engineer’s current capabilities. AI systems trained on domain-specific data — in this case, the structural calculations, MEP schedules, and design optimisation problems that professional engineers work through over careers — cannot be replicated quickly by a new entrant starting from a general-purpose language model. Consigli’s five years of development, in the specific context of AECOM’s engineering practice, represents a head start that is genuinely difficult to replicate from scratch.
Whether $390 million is the right price for that head start will depend on how the deployment goes. If the Autonomous Engineer can be integrated into AECOM’s project delivery at scale, the return on that investment could look very different from a standard software acquisition.
