- Start with a measurable cost. The AI use case worth starting with is the one tied to a cost you can already measure, like rework, delays or admin time. It isn’t necessarily the most impressive demo.
- Data decides more than the AI. If your models and project data aren’t consistent, the tool will struggle no matter how good it is.
- Write the return down first. A use case is only “worth it” if you can write down the expected return before you start, and check it afterwards.
- Quick wins come first. Early, visible results build the internal support you need for bigger AI projects later.
Most AEC firm leaders we talk to have the same problem. It isn’t a lack of AI ideas. It’s too many of them.
Every vendor has a use case. Every conference has a keynote. Someone on your team has already tried three tools on a free trial.
What’s missing is a clear way to decide which of these is worth your firm’s time and money.
There is real money at stake. Design errors and rework can eat up to 21% of a project’s turnover. Rework already costs the U.S. construction industry an estimated $88 billion a year, and much of it comes from bad or disconnected project data (Autodesk/FMI).
Meanwhile, construction productivity has grown about 1% a year for the last 20 years, while the wider economy grew almost three times faster.
So the opportunity is there. The hard part is picking the right place to start. Here’s the test we use with firms.
The worth-it test, before any pilot:
- Cost — Does it hit a cost you can already measure?
- Data — Is your data ready for it?
- Tools — Does it fit the tools you already use?
- People — Will your people actually use it?
- Return — Can you write the return down before you start?
- Risk — Does it protect your data and meet regulations?
A use case that can’t answer these yet isn’t ready for a pilot.
Test 1: Does it hit a cost you can already measure?
Start with the problem, not the tool.
A good AI use case points at a gap with a real cost: hours lost to rework, delays in review cycles, or time senior staff spend on admin and documentation.
If you can’t say roughly what the problem costs you today, you won’t be able to tell whether AI fixed it.
Questions to ask
- Where do we lose the most hours on rework or repeated checking?
- Which issues keep reaching clients or authorities when they should have been caught internally?
- What admin or documentation work eats into billable time?
QA/QC is often a strong first candidate for this reason. The cost of manual checking is visible, and firms using AI for QA/QC cut the cost of doing it by up to 20%.
Test 2: Is your data ready for it?
AI works on your data: BIM models, drawings, specs and project records.
If that data lives in silos, or model standards exist but aren’t applied consistently across projects, the AI will inherit the mess.
Before you commit, check the data it depends on
- Is it structured and consistent across projects and offices?
- Do your model standards get checked, or only written down?
- Is the infrastructure in place to keep it secure, whether cloud or on-premise?
A use case with a strong business case and poor data is usually a data project first and an AI project second. That’s fine, but budget for it.
Test 3: Does it fit the tools you already use?
A tool your team has to leave Revit to use, or that needs files exported and converted every time, will get used less than you expect.
Check how a use case fits your existing tech stack, and whether it can integrate with the systems your project teams already work in.
Test 4: Will your people actually use it?
Adoption is where many AI projects stall. Staff may be open to AI, but without a training path and clear ownership, tools sit unused.
This matters for retention too. Firms that still do QA/QC by hand are finding it harder to keep junior engineers and architects, who expect modern tools.
And the World Economic Forum reports that 40% of core skills for the average job will change in the next 5 years due to AI.
Questions to ask
- Who owns this tool day to day?
- What training does the team need?
- Which teams need to work together for it to succeed, such as design and QA/QC?
Test 5: Can you write the return down before you start?
This is the test most firms skip. Before a pilot, write down:
- The metric you expect to move: hours saved, issues caught, tender turnaround, revenue per employee.
- The baseline today.
- What result would count as success.
If budgeting for AI is ad hoc and there’s no cost-benefit analysis behind it, you’re relying on hope.
Tie every AI project to a specific revenue or productivity metric, and the “is it worth it” question answers itself.
Not sure where your firm should start? Book a free 30-minute consultation. We’ll talk through your workflows and which use cases look worth testing first.
Test 6: Does it protect your data and meet regulatory requirements?
BIM models hold sensitive design, cost and client data. Any AI use case has to protect that data and your IP.
Building authorities are also moving toward automated code-compliance checks. Firms that can’t produce clean, checkable project data risk slower approvals and more review rounds.
A use case that creates compliance or data-security risk can cost more than it saves.
Quick wins first, transformation second
Use cases fall roughly into two groups. Here’s how they differ.
| Quick wins | Long-term projects | |
|---|---|---|
| Scope | Efficiency gains in a single workflow, like QA/QC or documentation | Changes that cut across departments |
| What it needs | A workflow you already run | Better data, new processes and leadership alignment |
| When you see results | Within weeks | Sequenced on a 2–3 year roadmap |
Start with a quick win. Early results build support across the firm and justify the next investment. A 2–3 year roadmap then sequences the bigger moves.
How the VitruAI AI Readiness Audit answers this for your firm
Running these tests properly takes time and an outside view. That’s what the VitruAI AEC AI Readiness Audit is for.
We look at your firm across strategy and vision, data readiness, technology, people and processes, and financial and ROI feasibility.
How the work runs
- Initial consultation
- Interviews with principals, project leads, BIM managers and QA/QC staff
- Gap analysis benchmarked against how leading AEC firms use AI
- Presentation of findings
- A prioritized roadmap
What you get
- An ROI-focused executive summary with projected savings and quick wins for decision-makers.
- A detailed audit report with gap analysis and a roadmap with timelines and investment recommendations.
- A stakeholder workshop to walk leadership through the findings and next steps.
- Optional add-ons: deep compliance and regulation analysis, or a customized proof of concept that shows real ROI before a larger rollout.
Sized to your firm
Quick tips
- Write down what the problem costs you today before looking at any AI tool.
- Check your data before you check the AI.
- Pick one owner per use case.
- Define success in numbers before the pilot, not after.
- Start with a quick win that your team will see within weeks.
FAQ
What’s the best first AI use case for an architecture or engineering firm?
It depends on where your firm loses the most time and money. For many firms, QA/QC is a strong candidate because the cost of manual checking is easy to see and measure.
Do we need perfect data before we start with AI?
No, but you need to know how good it is. A readiness check shows which use cases your data can support now and what needs fixing first.
How do we measure whether an AI project paid off?
Pick one metric before you start, record the baseline, and compare after the pilot. Common metrics are hours saved, issues caught before external review, tender turnaround and revenue per employee.
What is an AI readiness audit?
It’s a structured review of your firm’s strategy, data, technology, people and finances, ending with a prioritized list of high-ROI AI opportunities and a roadmap to deliver them.
Is the audit only for large firms?
No. The Basic tier is designed for practices of up to about 50 staff.
Conclusion
You wouldn’t start a project without a fee proposal. An AI pilot deserves the same discipline.
The use case worth starting with is the one whose return you can write down today.
Here’s your first step: pick the problem that costs your firm the most hours, and write down roughly what it costs now. That number is the baseline every AI tool has to beat.
Find out which AI use case is worth it for your firm. Book a free 30-minute consultation. We’ll talk through your workflows and which audit tier fits your firm.
About VitruAI
VitruAI is the agentic AI layer for architecture and engineering: agents that review BIM models against building codes, standards and each firm’s own conventions, plus the audit and training that get a firm ready to run them. VitruAI is a spin-off of ADAIA, a UAE-based corporate AI consultancy, founded by Ian Arden, a venture investor and serial entrepreneur who has worked with AI since 2007.