How Arcadis scales its engineering expertise with Nomic

Every asset Arcadis designs passes through the same quiet bottleneck. Before a submittal is approved, a design is issued or a plant is commissioned, someone must read it against the standard it must meet: the specification, the client’s requirements, the calculations behind the drawings. On a large project, that means thousands of pages, checked page by page.

That is still the right standard. But it no longer scales. Deadlines are shorter, projects are more complex, and the senior engineers best placed to do the checking are in short supply, so their time becomes the bottleneck. In a firm of 34,000, the expertise is there but scaling it and connecting the right people to the right information, standards and project knowledge, when and where they need it, can be a challenge.

Arcadis did not set out to use AI for its own sake. Clients choose Arcadis for the knowledge, judgment and experience of its people, and that expertise remains at the heart of how it creates value. Its purpose is to improve quality of life, and that work depends on engineers having time for judgment: what a finding means, what the risk is, what to tell the client. The question was whether a tool could bring that knowledge into the workflows where teams design and decide, and give engineers time back across disciplines and countries, while working with files in established systems and keeping every finding traceable, reviewable and subject to an engineer’s professional judgment.

Nomic works across the files engineers already use, including PDF drawing sets and specifications, native DWG drawings and IFC models. It can locate, extract, compare and organize technical information, with every finding linked back to the relevant sheet, clause or page.

Nomic accelerates the first review by finding and cross-referencing the relevant evidence. The engineer verifies the findings, decides what matters and remains accountable for the final response. The workflows are built by domain experts: the engineers and designers who understand what a rigorous review looks like in their discipline. They test each workflow on real material, see where it needs refinement and improve it. One team, for example, continued refining its instructions until every requirement was returned word for word with a citation. More than 750 workflows now exist, with nearly 200 shared for other teams to adapt and reuse. The opportunity is to turn the strongest of these into repeatable review approaches that carry technical knowledge across projects, teams and geographies.

Use is tied to projects, not only to people. A project workspace is set up with an Arcadis job number, and its AI budget is approved by the project manager and the account manager. Clients are part of that decision too. Some do not yet allow AI on their work, and Arcadis respects that; where a client needs to agree first, that conversation comes before Nomic is used. Today, Nomic is in use on more than 40 percent of engagements, from water treatment plants and flood defenses to rail and power.

AI enablement is about people as much as technology. It means changing how engineers work. Arcadis made that change visible with a firm-wide leaderboard. Points reward the behaviors that matter most: using Nomic on live projects and reusing workflows that colleagues have already proven. Use on a live project scores more than activity alone, and running a colleague's workflow scores the same as building your own.

Adoption began with engineers identifying practical uses on live work, within governed access. Arcadis is now focusing that activity on priority sectors and on the workflows with the strongest potential to improve quality, delivery and client outcomes. Around 70% of activity on the platform involves checking technical material against a standard, specification or previous version. In AEC, the digital asset lifecycle is the information that follows an asset from first plan to decades of operation, passed between designers, engineers, contractors and operators. Nomic can be used at every stage, as Figure 1 shows.

Design and engineering is where most of the checking happens. Two examples show what that looks like on real projects.

Olga Anastasiu reviews structural calculations and contractor shop drawings on a large technology campus, where every connection has to be checked against the design. Instead of starting each package from scratch, she built a suite of 24 workflows, one for each kind of check. Her reviews took 67 hours instead of 185, and the quality held. “I can check every connection in detail,” she said. Fourteen of the workflows are now public, so structural engineers elsewhere in Arcadis start from checks that already work.

Wouter Leek, in Arcadis’ Mobility business, is responsible for a large delivery package designed by several teams across several systems, on a tight program. Reviewing every element manually would have required extensive effort across multiple sources and systems. However, by using Nomic, Wouter was able to locate and flag a discrepancy in the power-supply specification that could have led to the wrong materials being procured. Wouter then checked the evidence behind each comment before sharing the findings with the design teams, which updated their designs. “I verify if the comments made by Nomic are sensible and share the results with the teams,” he said.

Security was a condition, not a feature. Nomic is SOC 2 Type II audited and runs regional instances, including Europe, the UK and Australia, where contracts require it. Zero-retention agreements with the model providers mean Arcadis data never trains external models, and Nomic inherits existing access controls, so it never widens who can see what.

For the team that runs Nomic, the admin portal does the heavy lifting: SSO and SCIM provisioning, seats, spend limits and project budgets in one place, and usage by team, region and project.

Nomic also works with other agents. Through its Agent API, Arcadis developers launch Nomic workflows from their own tools. Arcadis is now exploring agent-to-agent connections through open standards such as the Model Context Protocol (MCP), so an agent drafting a report can ask Nomic what the specification says and get the citation back.

The partnership combines Nomic’s engineering-focused AI with Arcadis’ technical expertise, project knowledge and delivery scale. For example, Nomic’s forward-deployed engineers run weekly sessions with discipline groups on their own files, and what those sessions surface goes straight into the product. When an Arcadis tool hit a file size limit on the Agent API, the limit was raised.

In its 2027–2029 strategy, presented at its Capital Markets Day on September 29, 2026, Arcadis identified Nomic as an AI partner purpose-built for engineering. Arcadis also showed how Nomic has supported three winning technical tenders with combined contract values of more than €20 million. Across the firm, Nomic is now being used on client engagements worth more than €300 million, and engineers have identified 86 critical issues during review. The people closest to the work agree. Of the first eleven project managers reporting, five said the time went back into quality or scope, and ten would use Nomic again. Of 120 engineers surveyed, 82 percent find Nomic useful, and none of the 71 who answered on trust use its output unchecked.

Next, Arcadis moves to targeted programs in its largest sectors, with every result confirmed in writing by a named project manager. The aim continues to be to reduce the time engineers spend searching and cross-referencing information, so they can focus more of their expertise on technical assurance, managing risk and improving outcomes for clients.

Arcadis is the world’s leading company delivering data-driven sustainable design, engineering and consultancy solutions for natural and built assets. Its purpose is improving quality of life. Its 2027–2029 strategy, Arcadis Reset: Focused Growth, Sustainable Value, concentrates investment in transportation, energy and water, and industrial manufacturing and technology, with AI and digital embedded across the business.