AI Meets Infrastructure: The Rise of Open Engineering Workflows

The growing complexity and scale of infrastructure projects has created a pressing need for faster, more precise engineering workflows. With the AEC industry facing mounting workloads and workforce shortages, many organizations are implementing artificial intelligence (AI) to help tackle complex projects on tight schedules. Aided by AI, AEC organizations are automating repetitive tasks, increasing efficiency, and leveraging large-language models (LLMs) to develop and refine designs more quickly and precisely than ever before. 

Image source Jafree/stock.adobe.com.

The increasing complexity and scale of infrastructure projects has created a pressing need for faster, more precise engineering workflows. With the AEC industry facing mounting workloads and workforce shortages, many organizations are implementing artificial intelligence (AI) to help tackle complex projects on tight schedules. Aided by AI, AEC organizations are automating repetitive tasks, increasing efficiency, and leveraging large-language models (LLMs) to develop and refine designs more quickly and precisely than ever before.

To help meet the growing demand for technology-driven solutions, Bentley Systems has integrated AI into its infrastructure software portfolio, embedding LLMs grounded on engineering-specific knowledge, such as building codes and design standards. The convergence of infrastructure AI and an open, connected software ecosystem is transforming how engineers design, build, and operate infrastructure.

Infrastructure AI + Open Apps: Building the Digital Backbone

The growth of infrastructure AI was emphasized throughout Bentley Systems’ 2025 Year in Infrastructure and Going Digital Awards (YII), and AI interest has continued ramping up since YII. Bentley’s approach centers on combining open applications and actionable data to support AI-enabled workflows, throughout the infrastructure lifecycle — from design through construction and into operations.

In advance of last year’s YII, Bentley conducted a survey in collaboration with industry partners to gain insights into AI’s effect on the built environment. The team also prepared a white paper entitled “The Impact of Artificial Intelligence on the Built Environment.” When asked to what extent AI will impact current business models, 73% of respondents said they anticipate a significant or moderate impact on their current business model. For 24%, AI is already a major consideration, and they are taking active steps to adapt their model for significant future disruption.

Impact of AI on Current Business Model.
A vast majority of surveyed industry firms agree that AI impacts their current business model. Image source: “Artificial Intelligence – Impact on the Built Environment,” Pinsent Masons, Bentley Systems, Mott MacDonald, and Turner & Townsend, September 2025.

AI has been integrated in Bentley products such as OpenSite+, which employs generative AI for commercial, industrial, and residential site design. The software incorporates AI-assisted capabilities from tools such as the Bentley Copilot design assistant, including natural-language interactions, grading optimization, and automated drawing production to help streamline site design workflows. Working in conjunction with other industry standard tools such as GIS and CAD, OpenSite+ incorporates inputs such as building footprints and roadway designs, along with key design parameters, to integrate project elements and accelerate design and documentation.

OpenSite+ embeds LLMs trained on engineering-specific knowledge such as building codes and design parameters to accelerate site design. Click image to enlarge. Image source: Bentley Systems.

Bentley has also incorporated AI capabilities in OpenUtilities Substation+, a cloud-connected solution for designing electrical substations. With branching & syncing capabilities, Substation+ enables concurrent workflows with multiple designers, aids version control management, and provides schema-driven intelligence for faster modeling.

Other Bentley applications, such as MicroStation, OpenRoads Designer, and OpenRail Designer, are also being equipped with AI capabilities. Bentley Copilot has also been added to these applications, along with new AI tools that automatically organize labels on complex drawings, eliminating overlapping text and streamlining review and revision processes. In addition, other engineering tools like STAAD have been AI-enabled with Bentley Copilot capabilities.

The new AI-driven tools are bringing significant changes to AEC workflows. Turner & Townsend, a global consultancy that provides infrastructure project management and other services, has used AI to streamline numerous processes, according to Guy Beaumont, Digital Consulting Lead in the firm’s London office. Turner & Townsend was one of Bentley’s partners in conducting the AI survey and preparing the white paper.

“We're using AI to reform specific processes and activities within functions — within commercial, project controls, planning, and design,” said Beaumont in a keynote panel discussion on AI at YII. “We’re doing the same things better, faster with less inputs, making people more productive.”

As an example, Beaumont cited the firm’s airport work, where the firm has automated numerous processes. “We are delivering what we call a commercial intelligence platform. It seeks to automate and streamline some of those post-contract commercial administration management activities that take a lot of time.”

Mark Coates, Bentley Systems Vice President of Infrastructure Policy Advancement, led a YII panel discussion on AI with Anne-Marie Friel, Guy Beaumont, and YJ Kim. Image source: Bentley Systems.

The Critical Role of Actionable Data and Data Readiness

Beaumont and other panel members stressed that AI alone cannot modernize infrastructure; the industry needs to develop data readiness and actionable data – structured, contextualized, retrievable, and trusted data. This means AEC organizations need to properly prepare data for use in AI.

In developing data readiness, organizations need to focus on sound information management practices, such as data governance and data ethics, according to Anne-Marie Friel, Partner at Pinsent Masons, an international law firm that provides advisory services to the infrastructure sector, and also a Bentley partner on the survey.

“We need our results to be trustworthy. We need our results to be reliable. We also need our results to be fair,” noted Friel, adding that data readiness requires sometimes tedious data management. “It's doing the housekeeping. If you don't do that, it's like a house of cards. If you don't have a strong foundation, everything is going to cave in on itself.”

The fragmented nature of the AEC industry adds unique challenges to developing AI data readiness. With infrastructure specialists often working in “silos” using different processes, tools, and data formats, a strong focus on standardization and interoperability is required for successful use of AI, noted panelist YJ Kim, AI Technical Lead at AEC firm Mott MacDonald.

“Data is definitely the key element, along with process and standardization,” said Kim. Standardization can be aided by common data schemas to describe project information or organizational knowledge, added Kim, citing consistent tagging of information, such as documents, designs, and lessons learned across projects. “If we can bring the data in this fragmented environment and describe the data with a consistent format, then we can unlock the value of the data.”

Connecting AI with the Model Context Protocol (MCP) 

As organizations adopt AI for engineering workflows, data interoperability and contextual understanding remain significant challenges. Model Context Protocol (MCP) is an open-source framework designed to help AI systems better understand engineering data. Rather than interpreting model geometry alone, AI tools can use MCP to access information about how project elements relate to one another and how they function within a larger system. MCP also provides a standardized way for AI agents to interact with engineering applications, helping organizations connect tools and data more consistently across projects.

MCP’s rich context can improve the usefulness of AI-generated outputs by making project information more accessible and actionable. The protocol may also help organizations connect proprietary AI agents with engineering applications while maintaining established data governance and security practices. Bentley has introduced MCP integrations for applications such as STAAD.Pro and MicroStation, and has submitted those MCP-based connectors for Anthropic’s Claude platform.

"AI is incredibly powerful, but without deep engineering context and precision, it cannot deliver true value to infrastructure professionals," notes Francois Valois, Senior Vice President, Open Applications at Bentley Systems. "MCP is the missing link that provides this essential context, ensuring that as AEC firms build out their own agentic workflow automation, those agents can seamlessly and securely interact with the rich engineering data and engineering logic housed within our ecosystem."

AI in Practice: Accelerating Workflows and Enhancing Precision

Based on the Bentley survey, automation of document processes is the leading application of AI at infrastructure organizations, followed closely by optimization of design and engineering.

What does your organization currently use AI for?
Automation of document processes and optimization of design and engineering were the two leading AI uses cited by survey respondents. Image source: “Artificial Intelligence – Impact on the Built Environment,” Pinsent Masons, Bentley Systems, Mott MacDonald, and Turner & Townsend, September 2025.

The processes of managing documentation and extracting information from documents such as reports, specifications, and design information are well-suited to AI, according to Kim. “It's relatively smaller risk. And from a technical perspective, document automation largely relies on natural language processing, as well as small and large language models, and they're relatively mature,” she noted.

For more technical applications of AI, design and construction workflows can be seen in applications such as OpenSite+, Substation+, and SYNCHRO+, the next evolution of SYNCHRO 4D for construction modeling. By employing AI-driven automation to handle repetitive tasks, SYNCHRO+ enables teams to focus on more collaborative planning and estimating workflows.

SYNCHRO+ provides AI-driven automation capabilities for 4D construction planning. Click image to enlarge. Image source: Bentley Systems.

In addition to the AI capabilities implemented into Bentley applications, some AEC firms are using publicly available tools such as OpenAI’s ChatGPT, Anthropic’s Claude, or custom-built tools to develop designs, aid construction processes, and streamline operations and maintenance. As an example, asset management teams have been using AI along with digital twin technology to analyze pavement conditions. Using machine learning algorithms, teams have been able to more efficiently analyze images, pinpointing pavement defects with greater accuracy.

Regardless of the type of AI application, data quality is key to successful AI use, and organizations should maintain a clear audit trail of data input and output, noted Kim. “It's really important to have the audit trail. If we cannot verify and clearly show that transparency to our clients, we lose trust. In order to build that trust, having a dedicated audit trail will be very crucial.”

Auditability is also important in project management applications, such as cost estimation and productivity analysis, noted Beaumont. “[Extensive traceability logs have proven valuable in] financial operations and how much it costs to review a contract, for example, and summarize it,” he said. It's that financial traceability, what data it's pulling from and retrieving from different sources, and in what context it's used.”

In implementing AI, Turner & Townsend has been developing system tests to verify output, noted Beaumont.” Those kinds of tests need to be modeled and remodeled, as you use AI more and understand it more, to make sure they're still relevant and giving you the right results,” he said.

Future Work Environments: Open Data Ecosystems and Human-AI Collaboration

AI is clearly changing the work environment for infrastructure professionals. The extent of change remains to be seen, but two key elements are taking shape: open data ecosystems and human-AI collaboration.

The move toward open data ecosystems means less reliance on traditional CAD-based workflows and more use of tools with embedded LLMs such as OpenSite+ and Substation+, as well as increased AI capabilities in established tools such as OpenRoads. Instead of producing reams of paper plans and documentation, teams will rely more heavily on digital twins and models to design, build, and maintain projects. In a nutshell, open data ecosystems increase the potential for AI functionality by accessing broader datasets

Throughout YII, speakers noted that AI is not envisioned to replace engineers, but augment their capabilities. By handling repetitive, tedious tasks, AI can free up engineers for more critical thinking, such as considering multiple design alternatives. AI can also handle more complex modeling and simulation tasks in shorter timeframes, again freeing up engineers for more judgment-oriented tasks.

In another collaborative measure, Bentley announced the Infrastructure AI Co-Innovation Initiative to collaborate with industry professionals on AI-enhanced workflows. The initiative, open to Bentley users, is intended to examine how Bentley APIs can evolve to better support AI use cases and explore new commercial models that reflect the evolving balance between AI driven and humans. As part of the initiative, Bentley held a SYNCHRO+ Hackathon to explore how AI can expand the capabilities of SYNCHRO+.

The human-AI collaboration element is key to successful AI implementation, according to the panelists. Beaumont cited a need for continuous monitoring of outputs, as well as rules about when AI should return control to expert users for executing critical tasks. “Keeping humans in the loop is really important,” he noted.

Kim emphasized a need for continuous performance monitoring of AI by humans. “It's not about just doing this once, but we have to maintain the quality consistently before the results get actually consumed by the various downstream systems,” she noted.

Growing use of AI may also require new skill sets for infrastructure professionals. “We're not necessarily expecting project professionals to become data scientists,” noted Beaumont. But we are looking for them to become more data- and AI-literate.”

“You need people who understand how to drive collaboration,” added Friel. “People need to understand why they should be collaborating, how they should collaborate.”

Unlocking the Value of AI

As organizations and individuals become more familiar with AI, the capabilities of AI will become more tangible. While challenges must be addressed in areas such data readiness, data governance, output monitoring, and workforce training, the benefits of AI can transform an industry that has historically been slow to adopt new technologies.

With the entire infrastructure lifecycle becoming more connected via digital technology and actionable data, AI can help infrastructure professionals improve efficiency, productivity, decision-making and project outcomes. Ultimately this means higher-quality infrastructure projects and systems. Organizations adept with AI-related technologies will be those best equipped to unlock the full value of AI.

This article was sponsored by Bentley Systems.

***

 

 

 

Searching for more information about Architecture, Infrastructure, and Construction? 

Click here!