It’s well known that construction and mining projects carry heavy labor costs. This was plenty evident while we planned our approach to build for Plumbata an AI contract management system.
On every site, the same pattern stood out with labor forming the single biggest expense. Yet the contracts that governed it resisted consistent application. Collective bargaining agreements often ran over 600 pages. They used dense legal language that demanded expert review before anyone could confirm the right wages were paid.
The Plumbata team routinely had to bring in attorneys or outside specialists just to confirm whether contractor rates were legitimate. When markups ran 20 to 30 percent above union base rates, pushing back was an uphill battle.
The Plumbata team helps develop tools that deliver clear labor intelligence and track contract performance on construction sites. To build its contract management software, Plumbata reached out to GoML.
Why AI contract management software is not just about document scanning
Understanding why this problem has proven so difficult to fix helps explain what a real solution needs to look like. Take one example: the IBEW’s Ontario jurisdiction for a single trade.
This agreement alone contains more than 5,500 separate rules that control when and how labor gets paid. It also includes roughly 2,200 different hourly wage rates that shift depending on years of experience, location and shift type. Scale across multiple trades and several provinces, and the full scope of the challenge becomes clear. Previous tools focused mainly on scanning documents.
They pulled out the text but stopped short of interpreting what the agreements actually meant. When the Plumbata team reached out to Rishabh Sood, our founder and CEO, they both saw that simply digitizing the contracts would not be enough.
The greater value would come from turning those scanned pages into something practical a working model of the pay rules, thresholds, variations and conditions that could actually be used on site.
How GoML built Pactum an AI contract management software from unstructured PDFs
GoML used its Agentic AI accelerator to build and deploy Pactum, the first working version of Plumbata’s AI contract management engine. The platform accepts both plain text and PDF documents that need OCR. It pulls out 28 standard pieces of information from union agreements, including wage rates, worker classifications, geographic zones, overtime rules, shift conditions and effective dates.
Instead of forcing users to read hundreds of pages by hand, Pactum turns the agreements into clean, structured data that can be queried directly. Sitting on top of the extraction layer is a rules engine. It applies to the specific logic from each agreement to calculate overtime, adjust for different locations, and apply the right pay rates based on classification. This delivers consistent results across jurisdictions.
The system runs within the AWS ecosystem. It uses Amazon Bedrock with Claude Sonnet 3.7 for language processing, Amazon Textract for handling documents, PostgreSQL to store the structured data, and Amazon S3 to keep the original files. A React-based interface lets users upload agreements, review the extracted terms, and confirm their accuracy during this early phase.
Pactum connects to the rest of the Plumbata platform through RESTful APIs. This supports bulk processing and ties directly into field data capture and project analytics.
The picture of success
The numbers from the Plumbata case study show the impact of moving from manual contract review to AI-powered contract intelligence.
- 95% reduction in manual contract review time
What previously required attorneys and outside specialists now runs in minutes per agreement, allowing teams to review labor contracts at a fraction of the time and cost.
- 65% improvement in rate and rule interpretation accuracy
The system applies agreement logic consistently, reducing errors caused by manual interpretation of thousands of wage and classification rules.
- 60% faster preparation for labor cost analysis
Project teams can move from contract review to labor cost forecasting much faster, improving planning and budgeting decisions.
- 3× Improvement in Scalability Across Contracts and Jurisdictions
New agreements can be onboarded and analyzed without proportionally increasing staffing requirements, making expansion across regions significantly easier.
What used to take many days or even weeks now takes just minutes. The data continues to accumulate in value each time another agreement is created using Pactum.
As a result, it can be used for benchmarking purposes, cost analytics and other forms of data-driven analysis across projects and countries.
What working with GoML was like
GoML helped Plumbata build the business quickly and with less hassle.
"Without the expertise of GoML, we couldn’t reach this milestone at such an accelerated pace. We expect to scale significantly faster than we would have with our previous business model. Additionally to having a team of engineers focused on developing our solution, GoML helped us to move very quickly from identifying problems through having a working MVP without incurring additional delays or costs associated with creating our own internal technology team.”
- Bill McCormick, Founder at Plumbata
Why AI contract management software matters for construction and mining firms
One thing is clear from the Plumbata story is that manual contract review cannot keep up. With so many agreements, jurisdictions and changing pay rules in construction and mining, no team can review them by hand on a consistent basis. That inconsistency comes with a real price tag contractors have long used it as a cover for inflated markups, knowing that most clients don’t have the tools to challenge them effectively. And without good, clean, structured data to work from, there is no basis for the cost analytics or labor benchmarks that modern project management relies on.
Pactum addresses all three of these problems in one place. It structures the raw agreements, applies pay rules consistently across every contract that follows, and produces clean data that teams can use whether they're pricing upcoming work or reviewing what's already been awarded.
Looking to automate your use case with AI?
GoML helps companies like Plumbata turn complex, unstructured operational data into workable insights powered by our AI Matic framework. Whether you are dealing with contract management, compliance or field analytics, GoML's AI Data Analytics Accelerator provides the foundation for transforming unstructured operational data into actionable insights.
Explore GoML's AI Matic platform or get in touch for a demo.
