Kahua has introduced Noa, an embedded artificial-intelligence assistant for its construction project-management platform, aiming to bring search, content generation and no-code workflow design into the same system used to manage capital programs. The company announced the product May 13 from Atlanta and said Noa is available within the Kahua platform at launch.[1]
The release reflects a broader push by construction-software vendors to move AI from standalone chat tools and isolated features into core project systems. Kahua’s central pitch is governance: users should be able to work with project information without copying it into a separate AI service, while the assistant remains subject to the platform’s established permissions and accountability controls.[1]
Two modes: assistance and creation
Noa is built on Kahua AI, the company’s embedded intelligence layer, and is directed at owners, program managers and project-delivery teams handling complex capital programs. Its functions are divided into Noa Assists and Noa Creates.
Noa Assists is intended for everyday information work. Kahua says it can search and summarize project material, retrieve records, extract content, help draft written material, produce reports and support routine requests for information and submittals. Those capabilities target a persistent construction-management problem: locating the right information across large volumes of documents, records and project communications.[1]
Noa Creates extends the assistant into application and workflow configuration. Through natural-language prompts, users can design and deploy applications and workflows in kBuilder Canvas, Kahua’s AI-assisted no-code design environment. The company introduced kBuilder Canvas in February as a way to let users visualize, test and iterate on project-specific solutions through prompts or a visual builder rather than conventional programming or external development tools.[2]
The distinction matters because the latter feature could affect how organizations configure their project-management processes, not merely how employees query data. Kahua has described the offering as a tool for creating targeted workflows inside its platform; its announcement does not say that Noa independently makes high-risk construction, financial or scheduling decisions.[1]
Security is Kahua’s main differentiator
Kahua says Noa operates inside its existing environment, with project data governed by the platform’s access controls and security model. The company argues this eliminates the need to move sensitive project data into disconnected public AI tools, a concern for organizations managing contracts, drawings, invoices, specifications and other controlled records.[1]
The approach builds on Kahua’s earlier Enterprise AI architecture, which the company said uses private infrastructure for AI models and supports automated collection of data from invoices, equipment nameplates and specifications. That earlier announcement also described chat- and voice-based agentic data entry, conversational access to project information, and integration with existing permissions and governance controls.[3]
Kahua describes its broader platform as FedRAMP-authorized at the Moderate level and authorized for DoD Impact Level 2 workloads.[4] However, the May 13 Noa announcement does not specify the foundation model or models powering the assistant, model-provider relationships, inference architecture, retrieval approach, context limits, retention practices, audit-log design or human-approval requirements. It also provides no public benchmark results for answer quality, extraction accuracy or workflow-generation performance.
Construction AI moves into systems of record
The launch arrives as major construction-technology suppliers position AI as a native part of project platforms. Procore has promoted AI features spanning search, submittal generation and predictive analytics, while Autodesk has described Autodesk Assistant as a conversational interface evolving toward more agentic assistance across products including Construction Cloud.[5][6]
The opportunity is substantial, although vendor research should be interpreted cautiously. Procore-sponsored research reported that 18% of project time is lost searching for data and 28% is wasted through rework; 55% of construction leaders surveyed expected automation to disrupt the sector within five years.[5] These figures are not an independent assessment of Kahua or Noa, but they illustrate why document retrieval, reporting and process automation are prominent targets for software vendors.
Adoption remains early across the industry. RICS’s 2025 global report, based on responses from more than 2,200 construction professionals, found that about 45% reported no AI implementation and 34% were still running early pilots. Just under 12% reported regular use in specific processes, while 1.5% reported use across multiple processes. Skills shortages, systems integration, data availability and implementation costs were among the main obstacles cited.[7]
Kahua’s strategy is to address those barriers by putting AI in an established project-management information system and pairing it with no-code configuration. That remains a product-positioning claim rather than evidence that the approach delivers better results than separate AI tools or competing platforms.
Evidence of impact remains limited
In Kahua’s launch material, Verdantix analyst Sophie Planken-Bichler said construction AI is shifting away from standalone point products toward unified enterprise platforms and embedded intelligent workflows. She also warned that AI can increase, rather than reduce, fragmentation if organizations lack strong data foundations and governance.[1]
A customer reference from Ben Bohmann, associate director of design and construction at the University of Colorado Anschutz, said the approach reduced repetitive work while maintaining data security and could enable customized applications for project delivery. The statement is anecdotal: Kahua did not publish productivity measurements, a controlled comparison, customer-adoption figures or an independent case study with the Noa announcement.[1]
Likewise, Kahua’s February kBuilder Canvas release included a partner comment from K2 Consulting suggesting AI could reduce configuration cycles from weeks to minutes, but it was not presented as a published benchmark.[2] For prospective users, the practical questions will be how reliably Noa handles project-specific terminology and documents, what review controls apply to generated RFIs and submittals, and whether organizations can prevent inaccurate outputs or unauthorized workflow changes.
Noa therefore represents an expansion of Kahua’s product architecture rather than demonstrated proof of broad productivity gains. Its significance will depend on whether the company can translate its embedded-data and governance claims into reliable deployment at the scale required by large capital programs.
Editor’s Take
Kahua is making the right product-level bet: construction teams do not need another browser tab where people paste contracts, drawings and meeting notes into a general-purpose chatbot. They need AI where permissions, project records and workflows already live. If Noa can reliably retrieve the right version of a document and draft the first pass of an RFI or report without leaking data across projects, that alone can remove meaningful operational friction.
The more consequential feature is Noa Creates. Natural-language workflow configuration could give owner and program-management teams a faster way to adapt their systems to a project instead of waiting on developers or living with spreadsheet workarounds. But this is also where controls matter most. A generated workflow that routes an approval incorrectly is not a harmless hallucination; it can delay a payment, obscure accountability or create a contractual problem.
The launch announcement is light on the evidence buyers should demand: model details, retrieval accuracy, auditability, retention policies and measured customer outcomes. The next proof point is not a polished assistant demo; it is a deployed capital program showing that permission-aware answers, human review and no-code changes work consistently across messy real project data.
References
- Kahua, “Kahua Introduces Noa: Embedded AI in Construction Project Management Platform” – https://resources.kahua.com/news/kahua-introduces-noa-embedded-ai-in-construction-project-management-platform
- Kahua, “Kahua Introduces kBuilder Canvas: AI-Assisted No-Code Solution” – https://resources.kahua.com/news/kahua-introduces-kbuilder-canvas-ai-assisted-no-code-solution
- Kahua, “Kahua Launches Enterprise AI With a Security-First Approach” – https://resources.kahua.com/news/kahua-launches-enterprise-ai-with-a-security-first-approach
- Kahua, “Security” – https://www.kahua.com/security/?utm_source=openai
- Procore, “Future State of Construction Report” – https://www.procore.com/press/future-state-of-construction-report
- Autodesk, “The 3 Biggest Autodesk AI Takeaways From AU 2025” – https://adsknews.autodesk.com/en/news/the-3-biggest-autodesk-ai-takeaways-from-au-2025/
- RICS, “Artificial Intelligence in Construction Report” – https://www.rics.org/news-insights/artificial-intelligence-in-construction-report
