Nexus 1 is landing at a moment when builders are under pressure to do more with the data they already have without adding more systems, headcount or complexity. By embedding more than 25 AI agents directly into CMiC’s unified ERP, Nexus turns daily project activity, financial transactions, and field documentation into a conversational surface that project teams can actually use in real time.
Instead of treating AI as a bolt-on dashboard, CMiC has woven natural language, sentiment analysis, and document intelligence into the workflows contractors already rely on, from bank recs and cost code governance to drawing management and WIP reporting. The result is an intelligent business partner that speaks the language of construction and meets superintendents, project managers, and CFOs where they work.
Jeff Weiss, CRO of CMiC, talked with ERP Today about Nexus 1 and its future implications for the building and construction industries.
Question: Nexus is positioned as an intelligent business partner for builders. How are general contractors using natural language queries to make faster, better jobsite and project-level decisions?
Jeff Weiss: CMiC’s Nexus 1 represents a fundamental shift in how contractors interact with their project data. Rather than navigating complex menu structures or waiting for report packages, project managers and executives can now ask questions in plain language—”What’s my projected cost-to-complete on the downtown tower?” or “Which subcontractors are trending over budget this month?”—and receive immediate, data-grounded answers pulled directly from the live CMiC ERP.
What’s resonating most with GCs is the speed of insight. Decisions that previously required a finance analyst to pull a report, format it, and email it back—a cycle that could take hours or days—now happen in seconds at the point of need. Project executives are using CMiC’s Nexus in owner meetings to answer real-time questions about cost and schedule without leaving the room. Field leadership is surfacing RFI backlogs and submittal status without touching a desktop. The intelligence is embedded in the workflow, not bolted on as a separate tool.
Q: Nexus 1 uses more than 25 AI agents across construction financials and project controls. Where are you seeing the biggest impact on bank reconciliation, cost code maintenance and project cost analysis?
JW: The impact across those three areas is meaningfully different, which tells you something important about where AI creates the most value in construction finance.
On bank reconciliation, the Nexus agents are eliminating the most time-consuming manual matching work—flagging discrepancies, categorizing exceptions, and surfacing items that need human review rather than requiring a staff accountant to touch every transaction. For firms processing hundreds of subcontractor payments monthly, this is a material time savings.
Cost code maintenance has historically been one of the most under-appreciated data quality problems in construction accounting. Inconsistent coding decisions compound across jobs and make WIP reporting unreliable. Our agents can help normalize cost code application, flag anomalies, and suggest corrections before they propagate—essentially giving the accounting team a continuous quality control layer.
Project cost analysis is where the financial and operational worlds converge, and that’s where Nexus is generating the most executive attention. The ability to run scenario-based cost-at-completion projections, understand variance drivers, and get natural language explanations of what changed and why—that’s a capability GCs have wanted for years but couldn’t get without a dedicated analytics team.
Q: How does sentiment analysis in the Project Pulse Dashboard help construction executives spot schedule, safety or subcontractor risks on active jobs before they escalate in the field?
JW: Project Pulse is built on a core insight: The signals that precede a project problem almost always exist in the data before the problem becomes visible in the field. CMiC’s AI is continuously analyzing structured data such as project progress, cost trends, RFI aging, submittal cycle times, change order velocity and synthesizing those signals into an executive-readable risk picture.
The sentiment dimension comes from understanding directional momentum, not just point-in-time status. A job that’s 5% over budget but trending toward recovery looks very different from one that’s 5% over and deteriorating. Nexus surfaces that directionality explicitly so executives aren’t reading static snapshots.
For subcontractor risk specifically, the system can identify patterns that correlate with performance degradation—slow RFI responses, increasing change order frequency, payment disputes—and flag those relationships for PM attention before a defaulting sub creates a schedule crisis. The goal is converting reactive crisis management into proactive risk intervention, and the early adopters of Project Pulse are reporting exactly that shift in their executive review cadence.
Q: Contractors often struggle with messy project documents. How does Nexus’ drawing extraction, spec book organization, and submittal prefill reduce rework and RFIs on complex building projects?
JW: Document chaos is one of the most persistent—and expensive—problems in construction, and it’s largely a data extraction and organization problem that AI is well-suited to solve.
CMiC’s Nexus approaches this at multiple layers. On drawings, the AI is trained to recognize construction document conventions, extract relevant metadata, and link drawing content to the appropriate project elements within CMiC. This means a PM searching for a detail on a specific system gets a precise result rather than hunting through a file tree.
Spec book organization follows a similar logic—Nexus can parse Division structure, identify submittal requirements, and create the submittal log framework automatically rather than having a project engineer build it manually from scratch. That work typically takes days on a complex project; Nexus can do it in minutes with high accuracy.
Submittal prefill is where we’re seeing the most direct RFI reduction. When the system can pre-populate submittal packages with the relevant spec sections, required data fields, and responsible party assignments, the back-and-forth that generates RFIs shrinks dramatically. The quality of the first submission improves because the requirements were surfaced clearly at the start—not discovered through a rejection cycle.
Q: For construction finance teams, what changes when project accountants can generate job cost reports, WIP summaries and cash projections using natural language instead of manual reports or spreadsheets?
JW: What changes, fundamentally, is what construction finance professionals spend their time doing—and what they’re able to contribute to the business.
The manual report production cycle in construction accounting is consuming enormous capacity that could be directed toward analysis and advisory work. A project accountant who spends 60% of their week pulling, formatting, and distributing reports has limited capacity to actually interpret those reports and counsel project teams. Nexus changes that ratio dramatically.
With natural language report generation, a CFO can ask for a WIP roll-forward by division, compare it to last quarter, and get an explanation of the overbilling shiftin under a minute, without a staff accountant as an intermediary. Controller teams are generating cash position projections that account for scheduled billings, anticipated subcontractor payments and retention release timing. Analyses that previously required complex spreadsheet models maintained by one person who becomes a single point of failure.
The other dimension that gets less attention is auditability. When reports are generated from the live CMiC system rather than derived through a spreadsheet chain, there’s a clear, defensible data lineage. That matters at audit time and it matters when an owner disputes a billing position.
Q: From a field adoption standpoint, how does the modern Material 3 interface help superintendents, project managers, and back-office teams embrace AI-driven workflows on real construction jobs?
JW: Adoption is where most enterprise software investments go to die in construction, and CMiC’s investment in the Material 3 interface is a direct response to that reality.
The construction workforce spans a wide spectrum of technical fluency. A superintendent who’s been building for 30 years and a project engineer two years out of school interact with technology very differently. The Material 3 design system gives us a foundation that’s visually intuitive, responsive across device types, and consistent enough that users can develop muscle memory quickly.
For AI-driven workflows specifically, the interface presentation matters enormously. If the AI’s output looks like a system message that requires interpretation, adoption suffers. When CMiC’s Nexus surfaces a risk flag or a cost variance explanation in a format that’s readable and immediately actionable—with clear calls to action and drill-down paths—users engage with it rather than dismissing it.
We’ve also been deliberate about not forcing a single workflow model. Superintendents in the field interact primarily through mobile, with simplified task surfaces. Back-office teams have richer dashboard views. The AI layer is consistent underneath, but the presentation adapts to context. That flexibility is what makes enterprise-wide adoption achievable rather than theoretical.
Q: How do you tailor Nexus’ capabilities to construction, ensuring recommendations respect each contractor’s business rules, contracts and project delivery methods?
JW: This is one of the most important distinctions between Nexus and general-purpose AI tools that contractors might try to apply to their operations: Nexus is construction-native and client-aware.
General LLMs don’t understand GMP structures, retainage mechanics, AIA billing formats, or the difference between a unit-price and lump-sum contract. Nexus is trained on construction-specific data and domain logic, and it operates within the CMiC ERP environment where each client’s business rules, chart of accounts, and contract structures are already configured.
This means when a contractor asks Nexus to analyze cost-at-completion on a GMP project, the system understands the owner’s contingency structure, the buyout savings position, and the contract exposure—not just the raw cost numbers. When it surfaces a billing recommendation, it respects the approved schedule of values and the client’s specific retainage terms.
We also recognize that delivery method matters. Design-build projects surface risk differently than design-bid-build. CM-at-risk projects have different financial reporting obligations. Nexus is configured to understand which delivery context it’s operating in and adjust its analysis and recommendations accordingly. The goal is an AI that feels like it was built specifically for each contractor’s operation—because in a meaningful sense, it is.
Q: Across construction operations, where are you seeing the fastest time-to-value in project controls, financial management or analytics, and what ROI metrics are builders reporting after rolling out Nexus?
JW: The fastest time-to-value is consistently in financial management, specifically around the monthly reporting and billing cycle. Contractors who have historically spent the first 10 days of every month in a close cycle—pulling data, reconciling cost, building WIP schedules, formatting owner reports—are compressing that cycle significantly with Nexus. That’s tangible, measurable, and felt immediately by the finance team.
In project controls, the time-to-value curve is slightly longer because it’s tied to project duration and the accumulation of data that makes AI analysis meaningful. But by month two or three of active project tracking in Nexus, project teams are reporting that their weekly lookback conversations are fundamentally different—they’re spending less time establishing what happened and more time deciding what to do about it.
On ROI metrics, the numbers contractors are sharing with us cluster in a few categories: reduction in report preparation time (potentially 40 to 60%), faster identification of cost variances before they compound, reduction in audit-related labor due to improved data traceability, and in some cases, measurable improvement in billing accuracy that accelerates cash collection. The harder-to-quantify but frequently cited outcome is reduced executive anxiety—senior leaders feel more confident in their project portfolio visibility, and that has real organizational value.
Q: As part of CMiC’s multi-year AI roadmap, how do you see agentic workflows orchestrating more complex construction processes like change orders, pay applications, and supply chain coordination?
JW: The Nexus 1 release establishes the foundation—AI that can understand, analyze, and advise. The multi-year roadmap extends that toward AI that can act, not just inform, within defined workflow boundaries.
On change orders, the agentic vision is an AI that monitors the conditions that generate changes—design conflicts, owner-directed scope additions, differing site conditions—and initiates the documentation, pricing, and notification workflows automatically when those conditions are detected. The human decision points remain, but the administrative scaffolding around those decisions is automated. That’s meaningful because change order cycle time is one of the biggest cash flow variables for contractors.
Pay applications are a natural agentic workflow candidate because they’re highly structured and rule-governed. An agent that can validate lien waiver receipt, confirm stored material documentation, apply the correct retainage computation, and assemble the AIA G702/G703 package for PM review—eliminating the assembly labor while preserving the human approval step—that’s achievable in the near term.
Supply chain coordination is longer horizon but strategically important. The ability to monitor material commitments against schedule, flag procurement gaps before they affect critical path, and recommend resequencing when lead times shift—that’s where AI moves from financial intelligence to operational intelligence. CMiC’s position as a unified ERP for both financials and project controls gives us the data foundation to make that vision real.
Q: From large contractors versus regional builders, how does Nexus adapt to different construction business models and AI readiness levels without forcing a one-size-fits-all approach?
JW: Scalability isn’t just about transaction volume. It’s about organizational complexity and AI maturity, and Nexus is designed with that full spectrum in mind.
For ENR Top 400 contractors with sophisticated analytics teams, Nexus extends existing capabilities. These firms often have data scientists and BI resources; Nexus gives those teams a more capable foundation and reduces the volume of routine analytical work so they can focus on higher-order modeling. The integration with enterprise data environments, the configurability of AI agent behavior, and the depth of the financial and project controls feature set are optimized for that complexity.
For regional builders—contractors doing $50 to $300 million in revenue who don’t have dedicated analytics teams—Nexus is often their first meaningful exposure to AI-driven business intelligence. For these firms, the design priority is immediate usability without a lengthy onboarding runway. Natural language interaction removes the barrier of learning a reporting tool. Pre-configured dashboards surface the metrics that matter most for their business model. The AI meets them where they are rather than requiring them to develop new organizational capabilities before they can extract value.
The ACU-based consumption model we’ve introduced with Nexus also reflects this reality. It allows contractors to start with the capabilities most relevant to their current operations and scale usage as their AI fluency and organizational adoption grows. We’re not asking regional builders to commit to an enterprise analytics investment before they’ve experienced the value firsthand.
