AI Agent Operational Lift for Lee Mathews, A Cogent Company in Riverside, Missouri

Why AI matters at this scale

Lee Mathews operates in the 201–500 employee band, a sweet spot where the firm is large enough to have structured data and repeatable processes but small enough to lack a dedicated data science team. This mid-market size creates a unique AI opportunity: the ability to adopt off-the-shelf, verticalized AI tools that deliver immediate productivity gains without the overhead of custom model development. In the industrial engineering sector, particularly water infrastructure, margins depend on billable hour efficiency and winning competitive bids. AI can directly impact both.

The core business: engineering water's future

Founded in 1954 and based in Riverside, Missouri, Lee Mathews provides comprehensive mechanical, electrical, and structural engineering services. Their niche is water and wastewater treatment—designing plants, pump stations, and distribution systems for municipalities. This work is document-intensive, regulation-heavy, and reliant on specialized modeling software. The firm’s longevity suggests deep client relationships and a backlog of historical project data, which is the fuel for effective AI.

Three concrete AI opportunities with ROI

1. Generative design for capital projects. Water treatment plant design involves complex trade-offs between process efficiency, construction cost, and site constraints. AI-driven generative design tools can explore thousands of configurations in hours, outputting optimized P&IDs and 3D models. For a mid-sized firm, this reduces the preliminary engineering phase by 30–40%, allowing them to pursue more bids and win with technically superior, cost-optimized proposals. The ROI is measured in increased win rates and reduced senior engineer hours per bid.

2. Automated proposal and report generation. Engineers spend significant time writing technical specifications, feasibility studies, and RFP responses. A large language model (LLM) fine-tuned on the firm’s past deliverables and industry standards can produce first drafts in minutes. This frees up 10–15 hours per proposal, allowing engineers to focus on design review and client interaction. The payback period for such a tool is typically under six months, given the high cost of billable time.

3. Predictive asset analytics for clients. By applying machine learning to SCADA data and inspection records, Lee Mathews can offer new recurring revenue streams: predictive maintenance and condition assessment services for the plants they design. This shifts the business model from pure project-based consulting to long-term service contracts, improving revenue predictability and client stickiness.

Deployment risks specific to this size band

The primary risk is change management. A 200–500 person firm has established workflows and senior engineers who may distrust AI outputs. Mitigation requires starting with assistive tools that keep the engineer in the loop, not black-box automation. Data quality is another hurdle; historical project files may be unstructured and scattered across network drives. A dedicated data curation effort, even part-time, is essential. Finally, cybersecurity for public infrastructure data is paramount—any AI tool must meet strict client confidentiality requirements, favoring private cloud or on-premise deployment.