Provision launches Scope Agent for construction scoping

Provision has launched Scope Agent, an AI system for generating scope of work from drawings and specifications, and says it can reduce missed items while speeding up preconstruction. The company also published benchmark results showing Scope Agent outperforming Claude on a live plumbing scope test and winning customer validation from ProWest Constructors.

Why it matters: - Missing or unclear scope can drive costly rework in construction and expose general contractors to unexpected costs. - Provision is pitching Scope Agent as a way to make preconstruction faster, more consistent and less dependent on manual review. - The company says the product is built to help estimators catch missed items before bids go out, reducing downstream risk.

What happened: - Provision announced the general availability of Scope Agent, its AI system for automated scope-of-work generation from drawings and specifications. - The company also released a benchmark report comparing Scope Agent with Claude on plumbing scope extraction from a live construction project. - ProWest Constructors, a California general contractor and Scope Agent customer, said the tool cut time from award to subcontract writing by about 50%.

The details: - Provision’s internal validation dataset, based on human-prepared scope sheets from real projects across North America, showed 97.01% match accuracy on scope item extraction. - The same validation showed 96.02% subcontractor classification accuracy in the top 2 choices. - Provision said a professional estimator working to a fixed time budget reached 91.3% accuracy and needed about four days per project. - The new report, Plumbing Scope QA: AI Scope Agent vs. Claude, tested both systems on the same drawings from a live hospital project. - Scope Agent extracted 145 plumbing scope items at 91.7% verified accuracy. - Claude extracted 96 items at 72.9% accuracy. - Scope Agent captured 51% more scope items. - Scope Agent’s accuracy was 18.8 percentage points higher. - Of Claude’s 26 flagged rows, 18 included invented drawing marks, manufacturer model numbers or schedule items not present in the drawings. - Provision said Scope Agent produced zero fabricated items. - Provision attributes the performance gap to drawing-element identification, deep linking and estimator intelligence. - ProWest said Scope Agent saved a senior estimator three or four days on a county job in Santa Barbara and surfaced four or five missed items, including misread sheet metal requirements. - ProWest said scoping is both art and science, and the product gives the team a protocol and more consistency. - Scope Agent has also been used on live tenders with several of Canada’s largest general contractors, including EllisDon on the New Brunswick Museum project. - Provision said the output was used to issue a formal addendum that modified subcontractor scopes of work. - Provision believes this is the first documented case of AI-generated scope flowing directly into a bid package at a top-tier general contractor.

Between the lines: - Provision is making a broader case that general-purpose AI is not enough for construction estimating because fabrication and missed details can create liability. - The company is positioning Scope Agent as decision-support software for senior estimators, not a replacement for them. - The benchmark framing suggests Provision wants buyers to judge AI on trustworthiness and traceability, not just speed.

What's next: - Scope Agent is available now for general contractors, construction managers and trade contractors. - Teams can book a demo. - Provision is likely to keep using customer projects and benchmark comparisons to support adoption across North American preconstruction teams.

The bottom line: - Provision is trying to turn AI-assisted scoping into a practical bidroom tool: faster than manual review, more complete than general-purpose AI and specific enough to be used in live subcontract packages.