AI maturity stages for architecture and engineering firms
Five stages from first personal use of AI to AI connected to a firm's project and financial data, what each stage looks like in an architecture or engineering firm, and the work it takes to reach the next one.
The five stages
The stages measure what a firm has in place: a policy, measured workflows, and AI connected to firm data. A firm can hold enterprise licenses for every employee and still be at stage 2, because nobody has measured a workflow or connected the AI to firm data. The line between stage 2 and stage 3 is where we see most firms stop, and it is the same line MIT CISR found in its research.
MIT CISR research on maturity and financial performance
MIT's Center for Information Systems Research surveyed 721 enterprises in its 2022 Future Ready Survey and placed each in one of four stages of AI maturity. Enterprises in the first two stages, 62% of the total, had growth and profit below their industry average. Enterprises in stages 3 and 4 were above it.
The survey has two limits. It covers enterprises across industries rather than architecture and engineering firms specifically, and CISR notes that the results mainly describe analytical AI rather than generative or agentic AI. In our five stages, stage 3 is also the first one where a firm has a workflow timed before and after, so it is the first point where the savings can be shown. Our five stages split CISR's first stage into awareness and exploration, because that is where most firms are deciding their first policy and tools.
Reaching the next stage
Stage 1 to 2: start safely
Write an AI policy, approve a small set of tools, and train everyone on everyday use. The AI governance guide covers the policy, and a shadow AI survey shows which tools staff already rely on.
Stage 2 to 3: prove value
Pick two or three workflows, measure the time they take today, then build and measure pilots. Proposal drafting, RFI responses and searching past details and specifications are common first choices. The six stall points mostly show up here.
Stage 3 to 4: connect and scale
Connect the project archive, Revit, accounting and the CRM, and build agentic workflows with approval steps. Clean project and financial data decides how fast this goes, which AI implementation for architecture and engineering firms covers in detail.
Stage 4 to 5: predict and offer
Structure project data for forecasting, such as fee burn and staffing, and develop AI-based services for clients, such as post-occupancy analysis or design option studies.
At every stage, BCG's 10-20-70 approach applies: about 10% of the effort goes to algorithms, 20% to technology and data, and 70% to people and processes, meaning training, workflow redesign, roles and governance.
Assessing maturity by practice area
Practice areas in the same firm are usually at different stages. Business development may run a measured proposal workflow at stage 3, while design is at stage 1 and construction administration has no AI use yet. Rating each practice area separately shows where the next pilot fits and which areas need a policy and training first.
- Business development: Is there a written proposal workflow with a source folder and a reviewer?
- Design: Do designers use approved tools for concept images and precedent research?
- Documentation: Are any QA or code checks run with AI and measured against past sets?
- Construction administration: Are RFI and submittal responses drafted from specs and past answers?
- Operations: Can finance ask fee and staffing questions of clean, joined data?
Sources: Weill, Woerner and Sebastian, Enterprise AI Maturity Model, MIT CISR Research Briefing, December 2024, from the MIT CISR 2022 Future Ready Survey (N=721). BCG, AI at Scale and the 10-20-70 approach.
Frequently asked questions
What is an AI maturity model?
An AI maturity model describes the stages an organization goes through as it moves from individual experiments with AI to AI built into how it operates. Each stage names what the organization has in place, so a firm can see where it stands and what the next stage requires.
What stage are most architecture and engineering firms at?
Most firms we speak with are at stage 2: approved tools, sometimes an AI policy, and staff using AI for email, proposal text and renderings. MIT CISR found that 62% of enterprises across industries were in the first two stages of its four-stage model.
How long does it take to move from one stage to the next?
In our work with firms, moving from stage 1 to stage 2 takes one to three months for a policy, approved tools and training. Moving from stage 2 to stage 3 takes about a quarter per pilot workflow. Stage 4 depends on how clean the firm's project and financial data is, and usually takes a year or more.
How do we assess our firm's AI maturity?
Rate each practice area separately, such as business development, design, documentation, construction administration and operations, because they are rarely at the same stage. Then look at what each area lacks for the next stage: a policy, a measured pilot, connected data, or a reviewer.
Get the next AEC tech analysis in your inbox
Original research on AEC tools, AI adoption, and industry data. Free, irregular cadence, easy unsubscribe.
Need help applying this to your firm?
We advise AEC firms on technology and AI strategy. Start with a free 30-minute AI strategy session: your workflows, the right tools, and quick wins you can act on.
Book a free AI strategy session