Engineering Delivery, Augmented: AI-Assisted Design, Coordination and Quality

Building  project teams must interpret and verify large volumes of calculations, model information, coordination issues, and drawing content throughout project delivery. These activities require professional judgment, but they also include repetitive and high-volume tasks that consume time and can make  quality processes difficult to apply consistently across projects.

During preliminary design, engineering methods may be dispersed across spreadsheets, scripts, reference documents, and individual expertise. This  makes validated approaches difficult to reuse through different applications  and teams.

During multidisciplinary coordination, model-checking platforms can  generate extensive clash results. Identifying duplicates, grouping related  issues, establishing ownership, and determining which clashes require attention can demand substantial manual effort before technical coordination  even begins.

During drawing production and review, engineers and peer reviewers must  identify both technical concerns and repeatable documentation issues,  including missing references, inconsistent schedules, overlapping  information, and coordination discrepancies. When basic concerns are  discovered late, review cycles can become focused on corrections rather than  higher-value technical judgment.

The challenge was not to automate engineering responsibility. It was to determine where governed computation and artificial intelligence could  support project teams without creating an opaque or autonomous design  process.

WSP therefore began developing an integrated approach that combines  deterministic engineering services, AI-assisted issue interpretation, explainable drawing analysis, and human review. The objective is to move quality activities earlier in project delivery, support more consistent processes, and allow professionals to focus attention where experience and judgment create the greatest value.

WSP developed an AI-assisted project delivery framework that embeds support at  three connected stages of the engineering workflow: preliminary design,  multidisciplinary coordination, and deliverable review.

Structural APIs support design development and checking. Repeatable  engineering calculations developed by WSP engineer are converted into  documented computational services with defined inputs and outputs. Engineers and approved applications (including governed AI agents and chats) can use these services for preliminary design and verification, while future AI  interfaces can further orchestrate the workflow without relying on a language model to invent engineering calculations. Prototype documentation includes  interactive calculation examples, an API catalogue, and reference material  describing available calculations, required inputs, and returned results.

Clash Triage supports multidisciplinary coordination. The AI-assisted  platform groups, prioritizes, and streamlines issues produced through ACC  Model Coordination. It is intended to help teams convert high-volume clash  results into information that is more digestible and actionable, while  project professionals determine the required coordination response.

Redline supports drawing QA/QC. The tool combines deterministic checks with  AI-assisted analysis to identify potential drawing concerns and present  transparent findings, including supporting evidence, confidence information,  rule versioning, and review coverage. It provides an initial set of eyes  before formal review rather than replacing peer review or engineering  judgment.

Together, these capabilities shift assurance from a final checkpoint toward  a continuous process throughout design and coordination. AI handles  information volume and repetitive analysis; governed services perform defined  calculations; engineers validate findings and make decisions.