AI+AEC in News

AI agents and assistants: AI news

News on AI agents, assistants and large language models built for architecture, engineering and construction work, from drafting documents to working inside design models.

10 stories

AEC Magazine · 10 Nov 2026
Trimble embeds AI across AECO
Trimble Connect platform introduces AI agents and automation to improve project coordination, manage risk and streamline workflows Trimble has embedded AI within Trimble Connect , expanding the platform beyond its traditional role as a common data environment (CDE) to connect site information with design, cost and operational workflows. The new capabilities are intended to help teams identify and mitigate risks across the project lifecycle, while keeping human judgement at the centre of decision-making. Unveiled at Trimble’s Dimensions User Conference, the AI capabilities span design, construction and project operations. In design, they include generating architectural options using live as-built point-cloud data, converting natural-language prompts into constructible structural models, and automatically classifying scanned point clouds into structural elements. For construction teams, applications include AI-powered diagnostics for machine-control systems, converting 360-degree site walk videos into 3D point clouds, and voice-driven site inspections that automatically transcribe observations and link them to project drawings.
Summary and context →Read on AEC Magazine (opens in a new tab)
AEC Magazine · 10 Nov 2026
G2 Digital launches quad-GPU 2U server
UK-manufactured 2U Q-Power offers firms an on-premise alternative to premium cloud AI services G2 Digital, a UK specialist in custom-built rack-mount computers, has launched the 2U Q-Power, a high-performance GPU server designed for complex AI workloads and rendering. The system can accommodate up to four dual slot Nvidia RTX GPUs, offering twice the capacity of most conventional 2U servers and greater computing density within a rack. This is made possible by a custom chassis design that mounts the GPUs horizontally, parallel to the motherboard. When configured with four Nvidia RTX Pro 6000 Blackwell Max-Q GPUs, the Q-Power delivers 384 GB of combined GPU memory, enabling organisations to run demanding AI workloads entirely on-site. This includes large language models (LLMs) with 70 billion parameters at full precision, as well as significantly larger models using lower-precision formats and quantisation.
Summary and context →Read on AEC Magazine (opens in a new tab)
Archinect · 9 Nov 2026
AI companies are offering architects six figures for CAD files; have you been approached?
Artificial intelligence companies are turning to architects and other AEC firms in search of data. As tech enterprises seek to build AI models and agents to perform industry-specific tasks, such as reviewing drawings and writing RFI responses, architecture firms are being approached by businesses offering money in return for access to private data. “Architects, have you seen this?” asked USC adjunct professor Simon Ha in a LinkedIn post . “This AI company is offering to pay six figures for your CAD files, including revisions and markups, to train AI models on how design work gets done.” Ha’s post included a screenshot of a social media post by a company, Nyne.ai, describing itself as “the data factory for AI,” which said, “You’ve got 3 years of CAD files doing nothing. We’ll pay you six figures for them.” “This raises serious questions about copyright and ownership of instruments of service,” Ha added.
Summary and context →Read on Archinect (opens in a new tab)
AEC Magazine · 5 Nov 2026
Qonic AI launches inside BIM platform
AI agents work directly inside the model, letting users find, check, fix and create BIM data using plain language Qonic has introduced Qonic AI, a collection of agents built into its cloud-based, collaborative BIM modelling platform. Qonic AI works directly inside the BIM model, rather than sitting alongside it. Qonic has also launched an MCP server, which lets users connect their own AI assistant to the platform. Inside Qonic, through the “Ask Qonic” chatbot, users describe what they want in plain language, and Qonic’s agents find, check and update elements using the platform’s own tools. Tasks include completing missing properties, building reports and editing geometry. “Anything you can do in Qonic in terms of geometry and data creation, editing, and manipulation can be done through prompting,” said Qonic’s Jacob De Sutter.
Summary and context →Read on AEC Magazine (opens in a new tab)
AEC Magazine · 3 Nov 2026
Autom8Labs launches AEC skills for AI agents
Guides AI agent through tasks in Revit, AutoCAD, 3ds Max and more, with “clear steps” for doing the work and checking the result Autom8Labs has released 22 free MCP Skills, which work alongside the company’s relevant MCP Connectors for Revit, AutoCAD, Navisworks, Grasshopper, MicroStation, 3ds Max, Rhino, Civil 3D and Advance Steel. Rather than simply telling an AI agent what to do, each MCP skill guides it through the steps an experienced professional would follow. “Giving an AI agent access to a design application doesn’t automatically mean it knows the best way to work in it,” said James Gray, Autom8Labs. “The MCP Connectors give your AI agent the tools. Skills give it the method.”
Summary and context →Read on AEC Magazine (opens in a new tab)
AEC Magazine · 3 Nov 2026
nCircle Tech launches nVibe AI app builder
AEC professionals describe the workflow and AI application platform builds the app nCircle Tech has launched nVibe, an AI-powered development platform designed specifically to help AEC professionals turn ideas into working applications without needing coding skills. Domain experts, including site managers, project engineers and VDC leads, can describe a workflow in plain language, and nVibe builds the application within a sandboxed environment designed to give IT teams control over security, permissions and deployment. According to nCircle Tech, nVibe differs from general-purpose AI coding tools through its built-in understanding of AEC terminology and workflows. This includes concepts such as Work Breakdown Structure (WBS) codes, CSI MasterFormat, RFIs, submittals and LoD 100–500.
Summary and context →Read on AEC Magazine (opens in a new tab)
Construction Dive · 2 Nov 2026
Construction AI: Here are 5 questions to ask before you buy
It's not exactly a hot take to state that AI in construction is no longer a pilot project. In 2026, AI assistants on jobsites are answering questions in plain language and taking real steps on a contractor's behalf. This is quickly becoming the baseline across estimating, field operations, fleet management and all the reporting that comes with it. Many contractors now find themselves in an awkward spot. The upside is faster bids, fewer missed line items and less time spent retyping yesterday's logs. However, these same AI tools touch the numbers a company bids, bills and gets paid on, so caution is good business. Also, every vendor promises speed, because it's easy to demo, but trust is what protects margins. Just asking these five questions before pulling the trigger on an AI solution can drastically aid your evaluation process. 1. What data is it built on?
Summary and context →Read on Construction Dive (opens in a new tab)
AEC Magazine · 29 Oct 2026
Snaptrude Agent launches for BIM workflows
AI agent supports workflows from early concept design to detailed BIM modelling, structural systems, MEP and construction documentation. Snaptrude has introduced Snaptrude Agent, an AI agent for its BIM platform designed to support teams from early concept studies to construction-ready LoD 400 models incorporating structure, mechanical, electrical and plumbing (MEP), and drawings. Users are said to retain control of the process: they give the Agent a prompt or problem, review its proposed plan, then approve it or provide further direction. Discover what’s new in technology for architecture, engineering and construction — read the latest edition of AEC Magazine Subscribe FREE here According to Snaptrude, the Agent can automate tasks across the design process, including concept modelling, detailed BIM modelling and construction documentation. It can generate multiple massing options from a site analysis, ranging from a handful of alternatives to hundreds.
Summary and context →Read on AEC Magazine (opens in a new tab)
Construction Dive · 21 Oct 2026
Newforma acquires Lumi AI, will incorporate into its tech
Newforma, a construction information management technology provider, has acquired Lumi AI , an artificial intelligence startup, according to a Wednesday announcement. The terms of the transaction were not disclosed. With the move, Newforma will connect Toronto-based Lumi’s agentic AI into Vojo, Newforma’s AI assistant that lives on the company’s agentic framework, according to the announcement. Newforma will integrate the agent into its platforms that hold more than two decades of project data, which include RFIs, submittals and communication history. Lumi’s AI helps businesses use plain language queries to investigate and uncover patterns in their data. Going forward, planned capabilities for Vojo include natural language search and cross-project queries across project data, per the announcement. Newforma CEO Peter Cannone told Construction Dive via email that Lumi will not be accessible to non-Newforma customers.
Summary and context →Read on Construction Dive (opens in a new tab)
AEC Magazine · 11 Oct 2026
GPT-6 Astra and the future of BIM modelling
AI has made a significant leap forward in working with 3D geometry. Open AI’s latest model, GPT-6 Astra, can turn drawings, photographs and instructions into 3D models. For AEC, the bigger question is how this capability will fit into a design process in which almost everything is subject to change, writes Martyn Day There is something disconcerting about watching software attempt work that has occupied professionals for decades. Give Open AI’s new GPT-6 Astra model references of building, connect it to 3D modelling software and leave it to work, and geometry appears, sometimes looking more resolved than it is. Models once built object by object can now be attempted through conversation or uploaded drawings. Experiments by architects with ChatGPT Astra deserve attention. A general-purpose AI can interpret references, plan modelling operations, write code, inspect the result and make corrections.
Summary and context →Read on AEC Magazine (opens in a new tab)