Nomic ✓

Nomic is a New York City-based AI platform, founded in 2021 by Brandon Duderstadt and Andriy Mulyar, that has evolved into an AI workspace purpose-built for architecture, engineering, and construction (AEC) firms -- deploying domain-specific AI agents across drawings, specifications, BIM models, contracts, correspondence, schedules, submittals, and a firm's accumulated institutional knowledge. Rather than a single-purpose tool, Nomic positions itself as a horizontal AI layer spanning the design, engineering, construction, and facilities-management phases of a project's life.

The platform's agent capabilities are notably broad: Nomic's own materials describe handling 187 distinct tasks across construction administration, design, code compliance, drawing review, submittal processing, BIM modeling, CAD work, takeoffs, and operations. Concrete use cases include automated code compliance checking against more than 380 building codes, AI-assisted drawing review with direct PDF markups, submittal automation, firm-wide detail search across a company's own historical project archive, and project research.

A September 2026 announcement marked a significant validation moment for the company: Aurecon, an Asia-Pacific engineering consultancy with 6,700 employees, is scaling Nomic across its entire workforce following a four-year partnership that helped shape the platform's development. Simultaneously, Arcadis -- a global design and engineering consultancy -- announced it is embedding Nomic's AI agents into the core of its project-delivery model, following a six-month trial involving roughly 150 engineers across 12 countries. In that trial, 86% of participating engineers said Nomic changed how they approach their work, with the agents specifically analyzing building codes, design specifications, and technical documentation to flag risks and misalignments -- compressing what would otherwise be days of manual coordination work into agent processing plus expert human review.

Both Aurecon and Arcadis are now strategic investors in Nomic in addition to being commercial partners, a meaningful signal that two large, established AEC consultancies are betting institutional capital -- not just a software budget -- on the platform's continued development. Nomic's agents integrate with tools these firms already use operationally, including Autodesk Forma and Bentley ProjectWise, with further platform integrations described as in development.

Nomic's own framing describes its agents working "alongside people" to share and connect information beyond any single project -- a distinction worth noting, since it positions the platform's ambition as building firm-wide institutional memory and cross-project knowledge connections over time, not just automating tasks within the scope of one isolated project engagement.

Pricing is genuinely tiered and partly self-serve, a rarity among AI-agent platforms in this catalog: a Free tier requires no credit card and includes limited agent requests; Individual is $20/month with extended agent requests; Business is $40 per user/month (25-seat minimum) adding file integrations, team billing, SSO, and Agent API access; Enterprise is custom-priced with pooled usage commitments and dedicated support. AI usage itself functions as a consumable dollar balance across chat sessions, workflows, and document indexing, with Nomic's own documentation estimating $20 of usage covers roughly 80-200 Assistant sessions or 20-80 drawing-sheet reviews.

Editor’s takeBIM•STACK editor's take: Nomic's Aurecon and Arcadis relationships are the strongest evidence in this catalog's current AI-agent category that a platform has moved past pilot-stage enthusiasm into genuine, capital-backed institutional adoption -- two large global consultancies choosing to invest their own money alongside adopting the tool operationally is a meaningfully higher bar than a customer logo on a marketing page. The self-serve Free and Individual tiers also make Nomic one of the more accessible ways for a smaller firm to evaluate AI-agent workflows directly, rather than being funneled immediately into a sales-gated enterprise conversation.

Nomic's Assistant interface is built around a conversational, chat-based interaction model -- "ask me anything" against a firm's own project data -- which is generally quick to start using productively, since it doesn't require learning a new modeling or authoring paradigm, just how to phrase effective queries against the firm's indexed documents. Early sessions typically involve learning what kinds of questions the agents answer well versus where a human still needs to verify or dig deeper.

The more substantial organizational investment is in the Workflow and Agent API layers, where a firm builds out custom, repeatable automation (a specific QA check, a submittal-processing routine) tailored to its own standards -- that setup work compounds in value over time similarly to how a reusable skills library does in other AI-BIM platforms, but represents a genuine configuration investment beyond just chatting with the Assistant.

Arcadis's own six-month, 150-engineer, 12-country pilot is itself instructive on realistic learning-curve expectations at enterprise scale: a structured, multi-month trial across a genuinely large and geographically distributed engineering team, rather than a quick individual evaluation, is what it took to reach the reported 86% positive-sentiment result -- a useful benchmark for any large firm planning its own rollout timeline.