The architecture, engineering, and construction industry is a $13 trillion annual market that technology has largely bypassed. The materials and manufacturing firms that supply it face the same reality: ERP platforms from the 1980s, disconnected quality management systems, and operational data trapped in spreadsheets because the integration layer was never built.
The cost is measurable. Research by Autodesk and FMI found that construction professionals spend 35% of their time on non-productive activities[1] — searching for information, resolving conflicts, and managing rework. That rework costs the US construction industry $177 billion annually[1], with more than 70% traceable to design errors made before anyone broke ground. 85% of projects exceed budget; three quarters finish late.[1]
These are not execution failures. They are system failures — the predictable outcome of industries running software that was never designed for the complexity of the work it now supports.
Revit, acquired by Autodesk in 2002, holds over 95% market share[1] at major AEC firms globally. It is the industry system of record: taught in every architecture and engineering school, embedded in every firm's delivery workflows, and holding decades of project libraries in proprietary file formats. The same structural lock-in exists in materials — dominant MES and ERP platforms are deeply customized and operationally irreplaceable.
The instinct to build a better Revit or a better ERP is the wrong frame. Nobody replaced SAP by building a better ERP. They built around it until SAP became infrastructure no one touched. The opportunity in construction and materials is not displacement. It is activation — building the software layer that makes legacy systems productive for the first time.
Document review in AEC is a concrete example. Checking design documents for errors and coordination failures before a project goes to bid happens entirely outside Revit, takes three to six weeks, costs up to $100,000 per project, and catches only about 30% of issues that eventually become field change orders.[1] That is a $150B+ services workflow the incumbent left entirely unaddressed.[1]
Two conditions have converged. First, LLMs and vision models can now reason over the domain-specific, inconsistently structured data these industries produce. A BIM model contains enormous metadata — room classifications, occupancy loads, equipment specs, spatial relationships — that was previously unqueryable at scale. These models parse unstructured metadata, classify it semantically, and pass structured inputs to downstream engineering logic. That capability is new.
Second, demand pressure has made inaction untenable. AI infrastructure buildout is driving unprecedented MEP engineering volume. Housing and public infrastructure investment compounds the load. The engineering talent pipeline is not keeping pace. Firms are not evaluating AI because it is interesting. They are evaluating it because the manual model cannot absorb the project volume in front of them.
Building AI-native software for construction and materials firms. The highest-value targets are the workflows that legacy platforms never addressed: document intelligence that reads construction drawings and surfaces coordination failures; materials traceability systems that connect production data to quality outcomes; MEP design automation that ingests BIM models and generates compliant system layouts. These are not generic AI applications. They require deep domain encoding — building code logic, materials certification schemas, change order workflows — that takes significant investment to build correctly and creates durable defensibility once built.
Delivering faster and at lower cost through AI-driven engineering. AI coding assistants, automated test generation, and agentic delivery workflows are compressing the cost and timeline of custom software development. Forte Group has embedded these practices across its engineering teams. Requirements generated from meeting transcripts feed directly into story creation. Test suites are generated from acceptance criteria and validated before merge. This is not theoretical. It is the operational model the firm runs on today, and it directly affects the economics of every custom software engagement.
For construction and materials firms, this matters for two reasons. First, the software they need does not exist off the shelf — it must be built. Second, the firms that build it fastest and most reliably will define the competitive standard for their segment. Custom software built with AI-accelerated delivery is faster to market and cheaper to maintain than what was possible three years ago.
Building for these industries requires domain knowledge that cannot be shortcut. Understanding how ductwork routes around structural elements, how material certifications propagate through a supply chain, how a change order cascades across a project schedule — this context is not general. It must be earned through engagement with subject matter experts, and it determines whether an AI feature produces correct outputs or plausible-looking errors.
Data quality is the first obstacle in every engagement. The most valuable operational data in construction and materials firms often lives in PDFs, scanned documents, and legacy database schemas that predate any API layer. The integration and normalization work is typically the largest portion of the engagement and must be scoped honestly from the outset.
Compliance is not optional. A design error in construction is a liability event, not a software bug. AI outputs must be verifiable, auditable, and defensible under the regulatory frameworks these industries operate within. This requirement shapes architecture decisions from day one.
Forte Group has spent 25 years modernizing legacy technology for non-digital businesses. In construction and materials, that track record translates directly: the firm knows how to integrate with entrenched platforms, earn the trust of domain experts, build the data layer that makes legacy systems productive, and deliver AI-enabled software that meets the compliance and liability standards these industries require. Combined with AI-driven engineering practices that reduce delivery cost and compress timelines, Forte Group is positioned to enter this market at the right moment — and to build the kind of domain-specific software that creates lasting competitive value for the firms it serves.
The leading AI companies entering construction are not charging per seat for better design tools. They are pricing on outcomes: change orders prevented, margin recovered, schedule weeks saved. When AI expands a firm's capacity to take on work it would previously have turned away, the value created is structurally larger than what any legacy software budget line ever represented.
For construction and materials firms, the question is no longer whether AI is relevant. The $177 billion annual waste figure in US construction answers that. The question is whether the business will build the software layer to capture that value — or wait while a competitor does.
The built world is running on software from 1997. The firms that move now will set the standard for the next 25 years.
Sources
[1] Joe Schmidt IV, David Haber, et al. Every Building You've Ever Been In Was Designed By Software Built in 1997. a16z, March 30, 2026. https://www.a16z.news/p/every-building-youve-ever-been-in
Let's talk about where you are today and where you want to go - our experts are ready to help you move forward.
