There are three levels at which an AEC firm adopts AI, and treating them as one decision is why most firms stall. Tier 1 is individuals using tools on their own work. Tier 2 is a department wiring AI into a shared workflow. Tier 3 is the firm building AI into its infrastructure. They require different budgets, different owners, and different governance. Sort your AI questions into these three buckets and the roadmap stops being overwhelming.
"What's our AI strategy" is an unanswerable question because it's three questions wearing one coat. Here's how to take the coat off.
The AIA's 2025 report found 6% of architects using AI regularly and 8% of firms having implemented a solution. Bluebeam's 2026 outlook put broader AEC adoption at 27%, with 94% of those firms planning to expand next year.
Read those together and the story is uneven adoption, not slow adoption. Some firms are compounding while most haven't started. The firms that are compounding tend to have one thing in common: they stopped asking "should we use AI" and started asking "at which level, for what, owned by whom." That's the framework below.
Why "our AI strategy" is the wrong unit
When a principal says the firm needs an AI strategy, three different problems are hiding inside that sentence.
One is a designer who wants to use an AI tool to speed up a first-draft narrative. Another is the BD team wanting a repeatable, shared way to analyze incoming RFPs. A third is the managing partner wondering whether AI belongs in the firm's core systems, its DAM, its CRM, its project delivery.
Those aren't the same decision. The first costs $20 a month and needs a usage policy. The third is a multi-year infrastructure question with real budget and real risk. Bundle them and you get paralysis, because the risk profile of the biggest one freezes the smallest one.
Separate them into tiers and each becomes tractable.
Tier 1: Personal
What it is. Individuals using AI tools on their own tasks. A project architect drafting an email. A marketer summarizing a debrief. Someone pasting a scope into a chatbot to get a first-pass outline.
What it costs. Almost nothing. Consumer and team subscriptions run $20 to $30 per user per month. This is the cheapest, fastest, lowest-risk tier, and it's where nearly all firms actually start whether leadership planned it or not.
Who owns it. The individual, inside a policy leadership sets. This is the key move at Tier 1: your people are already here. The only real decision is whether they're operating inside guardrails or making it up alone.
The trap. Assuming personal-tier wins scale to the firm. They don't. One person getting 30 minutes back on emails is real, but it's invisible on the firm's P&L and it doesn't compound. Tier 1 is where you build literacy and comfort, not where you get transformation. Treat it as the on-ramp.
The Tier 1 governance minimum. A one-page policy: which tools are approved, what data must never be pasted into a public model (client-confidential material, anything under NDA, personnel information), and who to ask when unsure. That single page prevents most of the "shadow AI" risk that Bluebeam's respondents worry about, where 42% flagged data security as their top concern.
Tier 2: Department
What it is. A team wiring AI into a shared, repeatable workflow. Not one person's habit, a process the whole department runs the same way. This is where AEC firms get their first defensible ROI.
Examples that work. The BD team running every incoming RFP through the same AI-assisted analysis and go/no-go. Marketing generating proposal first drafts from a connected content library. Project teams turning meeting recordings into structured minutes and action items on a standard cadence. Bluebeam's early adopters concentrated their gains in exactly these areas: scheduling, planning, and document analysis.
What it costs. More than Tier 1, less than Tier 3. Purpose-built tools for a department's job, typically priced per seat or per team, plus the real cost, which is the time to design the workflow and get the team to actually adopt it.
Who owns it. A department lead, with a defined workflow and a data connection. The difference from Tier 1 is that the AI is connected to shared firm data, not just an individual's clipboard. That connection is what turns generic output into useful output, and it's what makes the win survive when the one enthusiast goes on vacation.
Why this is the tier that matters most. Department-level workflows are where AI stops being a personal productivity toy and starts moving a number the firm reports on: win rate, proposal turnaround, utilization. It's ambitious enough to matter and contained enough to actually finish. If your firm does one thing this year, make it a Tier 2 workflow in the department where the bottleneck is worst.
Tier 3: Organizational
What it is. AI built into the firm's core infrastructure and delivery. Firm-wide platforms, deep integration with your systems of record, AI touching project delivery itself.
What it costs. The most, in money, time, and risk. This is a multi-quarter or multi-year commitment with IT ownership, security review, and change management across the whole firm.
Who owns it. Firm leadership and IT, together. It needs a real budget line, an executive sponsor, and a governance structure, because at this tier a bad decision is expensive to unwind.
The honest caution. Very few firms should start here. Jumping to Tier 3 before you've proven value at Tier 2 is how firms buy expensive platforms nobody uses. Earn the right to Tier 3 by succeeding at Tier 2 first. The department wins teach you what the firm actually needs, and they build the internal credibility that a firm-wide rollout requires.
How to sequence it
The tiers are a ladder, not a menu. Climb in order.
Now: Set the Tier 1 policy. Your people are already using AI. Give them one page of guardrails and a short list of approved tools this week. It costs nothing and closes your biggest current risk.
This quarter: Pick one Tier 2 workflow. Choose the department with the most painful, most repeatable bottleneck. For most AEC firms that's business development, because pursuit work is high-stakes, time-hungry, and already standardized enough to systematize. Design the workflow, connect it to real data, and get the team using it the same way every time.
When Tier 2 has proven out: Consider Tier 3. Let the department wins tell you where firm-wide infrastructure is worth the spend. Don't guess at it from the top.
The firms compounding on AI got there by making their bets in the right order. Sequence beat size.
Where Foveate fits
Business development is the Tier 2 workflow we built for. Foveate gives a BD team one shared, repeatable system: RFP research matched to your real portfolio, go/no-go analysis, stakeholder prep for the people who decide, client-specific interactive presentations, and tracking on what happens after you send. It's connected to your firm's actual project history and staff, which is what makes it a department system instead of a personal tool.
If you're deciding where to place your first serious AI workflow, pursuits are usually the best place to start. Book a demo and we'll walk through what a Tier 2 BD workflow looks like in practice.
Frequently Asked Questions
Which tier should our firm start with? Set the Tier 1 policy immediately since your staff are already using AI, then invest your real effort in one Tier 2 department workflow this quarter. Skip straight to Tier 3 only after a department-level win has proven the value.
How much should a firm budget for AI? It depends entirely on the tier. Tier 1 is roughly $20 to $30 per user per month. Tier 2 is per-seat or per-team tool cost plus workflow design time. Tier 3 is a multi-quarter infrastructure budget with IT ownership. Budgeting for AI as one line item is the mistake the framework prevents.
What's the difference between a personal AI tool and a department AI workflow? A personal tool helps one individual on their own tasks and lives on their clipboard. A department workflow is connected to shared firm data, run the same way by the whole team, and tied to a metric the firm reports on. The data connection and the repeatability are what separate the two.
Why not roll out AI firm-wide from the start? Because firm-wide rollouts before proven department value tend to become expensive platforms nobody adopts. Department wins reveal what the firm actually needs and build the internal credibility a full rollout requires. Sequence matters more than scale.
Where do most AEC firms get the fastest ROI? In Tier 2 document-heavy workflows: proposal and RFP analysis, meeting documentation, and scheduling. Bluebeam's early adopters concentrated their reported savings in document analysis, scheduling, and planning.
Sources
About the Author
Kitae Kim
Architect with 10 years of experience in design and client communication. Co-founder of Foveate, the Pursuit Intelligence Platform for AEC firms. Former studio lead who saw too many winning designs lose to worse proposals.