AI Agent Operational Lift for Crawford, Murphy & Tilly in Springfield, Illinois

Why AI matters at this scale

Crawford, Murphy & Tilly (CMT) is a 75-year-old civil engineering firm headquartered in Springfield, Illinois, with 201-500 employees. The firm specializes in transportation, water resources, aviation, and site development projects for public and private clients across the Midwest. At this size—large enough to have deep project archives but small enough to pivot quickly—CMT sits in a sweet spot for AI adoption. The firm lacks the bureaucratic inertia of mega-engineering corporations, yet has the project volume and data density to make machine learning models statistically meaningful.

The civil engineering sector has historically lagged in digital transformation, relying heavily on manual CAD drafting, spreadsheet-based calculations, and paper-driven field reports. This creates a significant opportunity for a mid-market firm like CMT to leapfrog competitors by embedding AI into its core workflows. With billable rates under constant pressure and a tight labor market for experienced engineers, AI-driven productivity gains directly translate to margin improvement and faster project delivery.

Three concrete AI opportunities with ROI framing

1. Generative design for site development. Conceptual site layout—balancing grading, utilities, stormwater, and access—is iterative and time-intensive. By training a generative adversarial network on CMT's decades of past site plans, the firm can produce code-compliant layout options in minutes. Assuming a senior engineer spends 40 hours per concept at $200/hr, reducing that by 30% saves $2,400 per project. Across 100 site projects annually, that's $240,000 in recovered billable capacity.

2. Predictive analytics for project risk. CMT can build a model using historical project data (final cost vs. estimate, schedule variance, change order frequency) to predict which active projects are likely to overrun. Flagging a $5M project with a 10% cost overrun risk early allows for mitigation that could save $500,000. Even a 20% reduction in overruns across a portfolio of 50 active projects yields substantial ROI.

3. LLM-powered proposal automation. Responding to RFPs for municipal and DOT contracts is a major overhead. A retrieval-augmented generation (RAG) system can draft technical approach sections by pulling from past winning proposals, staff resumes, and project sheets. If a proposal manager saves 20 hours per pursuit at a blended rate of $150/hr, and CMT pursues 60 major contracts yearly, the annual savings approach $180,000 while improving response quality.

Deployment risks specific to this size band

Mid-market firms face unique AI adoption challenges. First, CMT likely has data scattered across network drives, project-specific folders, and legacy systems like Deltek Vision. Cleaning and centralizing this data is a prerequisite that requires dedicated IT effort. Second, the firm may lack in-house data science talent, making a hybrid approach—partnering with an AI consultancy for model development while training internal champions—the most practical path. Third, professional liability concerns mean any AI-generated design output must have a clear human-in-the-loop review process, documented to satisfy insurers and state licensing boards. Finally, change management is critical: veteran engineers may distrust black-box recommendations. Starting with assistive tools that augment rather than replace their judgment will drive adoption. A phased rollout, beginning with internal analytics and gradually moving to client-facing deliverables, balances innovation with the firm's reputation for reliability.