AI for Architects: The 2026 Founder's Playbook
AI for Architects Is Not a Software Question. It Is a Business One
Here is the number that should keep every principal awake at night: as of the American Institute of Architects 2025 research, only about 8 percent of architecture firms have actually implemented AI into their practice, with roughly another 20 percent still working on it. That means AI for architects, right now, is a market where the vast majority of your competitors are standing still. I run companies for a living. I have watched enough industries reprice themselves in real time to tell you what that gap really is. It is not a technology gap. It is a two-year head start sitting on the table, unclaimed, in the most fragmented professional services market in the built world.
I am a founder, not a consultant. I have built and operated businesses across hospitality, sport, healthcare adjacent services, and marketing, and I have watched the same pattern repeat every single time. The firms that treated a new technology as a tool got a modest bump. The firms that treated it as an operating model change repriced their entire economics. Artificial intelligence for architects is the exact same fork in the road, arriving right now, and most of the profession is walking down the wrong branch of it.
This is not another list of AI tools. You do not need me to tell you that Midjourney renders and that a chatbot can draft an email. You need someone who has actually moved revenue with these mechanisms, in real companies, to tell you where the money is, what to ignore, and how to sequence the change so it does not blow up your firm. That is what this playbook is.
Why I Am Writing This for Architects Specifically
Architecture is a business that hides its own economics from the people running it. You sell time. You bill against phases. You win work through relationships and reputation, and you lose money on the parts of the job nobody wants to talk about: the redraws, the coordination, the proposals you did not win, the utilization gaps between projects. AI touches every single one of those pressure points. Not the glamorous rendering part. The boring, expensive, margin-destroying parts.
That is the lens I am going to use for the entire article. Not "look what AI can draw." Instead: where does AI change the unit economics of running an architecture firm, and how do you capture that before your competitors do.
The Real State of AI for Architecture Firms in 2026
Let me give you the honest landscape, backed by real sources, because most of what you read on this topic is either vendor hype or academic hand-wringing.
At the macro level, the McKinsey State of AI research shows that the overwhelming majority of organizations now report using AI in at least one function. But here is the trap inside that headline. McKinsey also finds that only a small fraction of companies, roughly one in three, have actually scaled AI across the organization. The rest are stuck in what the report bluntly calls pilot purgatory. Adoption is easy. Transformation is rare. That distinction is the whole game.
Now zoom into architecture specifically. The AIA research is stark: a single-digit percentage of firms have genuinely implemented AI, and the movement is led disproportionately by large firms of fifty-plus people. Meanwhile industry surveys from the profession suggest that around half of architecture professionals now say they use AI tools in some capacity, with figures climbing fast year over year. Reconcile those two facts and you get the truth: a lot of individual architects are experimenting with AI, but very few firms have turned that experimentation into a system that changes how the business makes money.
Adoption Versus Transformation
This is the single most important idea in the entire piece, so I want to slow down on it.
- Adoption is an architect using a chatbot to reword a scope narrative. It is one person, saving one hour, invisibly, with no change to how the firm operates.
- Transformation is your firm redesigning the entire proposal-to-contract workflow so that a bid that used to take two weeks now takes three days, with a higher win rate, tracked and repeatable across every project lead.
The first is a productivity hack. The second reprices your business development function. McKinsey's own data underlines this: the firms getting real financial return are the ones that redesigned workflows, not the ones that bolted a tool onto an old process. I have lived this. In every company I have turned around, the win came from redesigning the process around the new capability, never from sprinkling the capability on top of a broken one.
If you take one thing from this article: do not ask "which AI tool should my architecture firm buy." Ask "which of my firm's workflows is most expensive and most repetitive, and how do I rebuild it around AI." The tool is downstream of that decision, and it barely matters.
Why the Slow Adoption Is Actually Your Opportunity
The architecture, engineering, and construction sector has a well-earned reputation for being slow to adopt new technology. Most people in the industry treat that as a problem. I treat it as the single biggest arbitrage available to a sharp principal right now.
When an entire sector moves slowly, the early mover does not get a 10 percent edge. They get a structural one. They win bids faster, so they bid on more work. They deliver documentation faster, so they take on more projects with the same headcount. They market more consistently, so their pipeline compounds. By the time the laggards catch up in three years, the early movers have already used those extra wins to hire, to build reputation, and to lock in the client relationships that matter. Slow sectors do not reward patience. They punish it.
Artificial Intelligence for Architects: Where the Money Actually Is
Now let me get concrete. I am going to map the mechanisms I have personally used to move revenue and margin in other businesses directly onto an architecture firm. These are not hypotheticals. Each one is a mechanism I have watched produce real numbers, transposed into your world.
Before the case studies, here is the map of where AI creates value in an architecture practice. Notice that design rendering, the thing everyone talks about, is not even the highest-leverage item.
| AI Lever | What It Actually Does | Business Impact | Difficulty | |
|---|---|---|---|---|
| Proposal and bid generation | Drafts scopes, fee narratives, and RFP responses from your past work | Higher win rate, more bids submitted | Low | |
| Client acquisition and marketing | Automates lead nurture, content, and outreach | Fuller top of funnel | Low | |
| Documentation and spec automation | Generates and cross-checks specs against drawings | Fewer errors, faster CDs | Medium | |
| Design iteration and massing | Produces early options from site and program data | Faster concept phase, more options shown | Medium | |
| Back-office automation | Automates admin, billing prep, project reporting | Reclaimed billable hours | Low | |
| Capacity and utilization planning | Predicts staffing needs and gaps across the pipeline | Higher utilization, better margins | High |
Let me walk through the four mechanisms that map most directly onto results I have delivered.
Winning More Bids: The Proposal Mechanism
I once worked with a hotel that lifted revenue from roughly nine million to ten million by using predictive tools for pricing and capacity management. On the surface that has nothing to do with architecture. Underneath, the mechanism is identical: use data you already have to make better commercial decisions faster than the competition.
For an architecture firm, the equivalent is the proposal and bid process. AEC firms routinely spend, on average, more than two weeks responding to a single RFP, and a meaningful share take twenty days or more. That is time, at senior rates, spent on work you might not win. Industry data on proposal teams that adopted generative AI shows response times falling by a third or more, with measurable improvements in win rates.
Transpose the hotel mechanism. Instead of predicting room demand, you build a system that ingests your entire history of past proposals, won and lost, and drafts new responses in your voice, pre-populated with your relevant project experience, fee logic, and boilerplate that used to eat a principal's weekend. The partner reviews and sharpens instead of writing from a blank page. You bid on more work because each bid costs you less. You win more because each bid is sharper and submitted faster. That is the same lever that took a hotel from nine to ten million, pointed at your business development pipeline. I unpack the deeper commercial version of this in my guide to AI for professional services, because architecture shares the exact economics of every other expertise-for-hire business.
More Clients: The Marketing Mechanism
I helped a sports business, WSB Sport, lift sales by around 30 percent using AI-driven marketing, and separately helped an agritourism business roughly double its guests through the same category of mechanism. Different industries, identical engine: AI systematizing the top of the funnel so that lead generation stops depending on whoever happens to have time that week.
Most architecture firms market like it is 2005. A dormant website, a portfolio that has not moved in a year, and new work that arrives entirely through referral and luck. That is not a marketing strategy. That is hoping.
The mechanism that doubled an agritourism's guests transposes cleanly: AI-generated content built around your real project expertise, published consistently; automated nurture sequences that keep your firm in front of past clients and prospects; outreach that is personalized at scale instead of generic and rare. You are not trying to go viral. You are trying to make sure that when a developer, a facilities director, or a homeowner is ready to commission work, your firm is the one already in their inbox. I go deep on how to build this engine in my AI marketing strategy frameworks and tools piece, and the client-acquisition specifics in the AI for sales guide.
More Output From the Same Team: The Capacity Mechanism
The result I am proudest of came from a medical center where we lifted operational capacity by about 20 percent, not by adding staff, but by using AI to optimize scheduling and patient flows. Read that sentence again through an architect's eyes. Twenty percent more throughput from the same people, purely by fixing how work moved through the system.
An architecture firm is a scheduling and flow problem wearing a design costume. Your margin is destroyed in the gaps: the drafter idle between projects, the senior architect doing coordination a system could catch, the CAD hours spent on documentation that repeats across jobs. The medical center mechanism transposes directly: use AI to predict staffing needs across your pipeline, to catch documentation and coordination errors before they cost a redraw, and to automate the repetitive drawing and spec work that has no business consuming a licensed professional's time.
Twenty percent more capacity without hiring, in a business that sells capacity, is not an efficiency story. It is a margin story. It flows straight to the bottom line or lets you take on more work with the team you already have. The nuts and bolts of building these systems live in my AI workflow automation business guide.
Faster, Cheaper Delivery: The Documentation Mechanism
The least glamorous lever is often the most profitable. In the AIA research, architects themselves flagged complex specifications, cost estimation, updating product lists, and coordinating specs with drawings as the areas with the most room for AI-driven improvement. These are the tasks nobody markets and everybody dreads, and they are exactly where quiet money hides.
Generative AI can draft specifications, cross-check them against drawings to flag inconsistencies, and keep product data current, turning a multi-day slog into a supervised review. Tools from established platforms like Autodesk's AI offerings are pushing this into mainstream BIM workflows. The point is not the tool. The point is that every hour of senior time you claw back from documentation is an hour redeployed to design, to clients, or to bids you actually win.
The Design Iteration Mechanism, Handled Honestly
I have deliberately left design last, because it is where the hype is loudest and the business case is softest. Yes, generative tools can now take site data and program requirements and spit out early massing options in minutes rather than the day or two of manual study they used to require. Early published examples from the profession show exactly this: a system extracting site and program data automatically and producing massing options that previously took days of research.
Here is the honest read. This is genuinely useful at the concept phase, because it lets you show a client more options faster and explore a wider solution space before committing. But it is not where your firm's money is hiding, and it is not where I would tell you to start. Design iteration is the dessert. Proposals, documentation, marketing, and capacity are the meal. Show a client three AI-assisted concept options and you have impressed them once. Rebuild your proposal engine and you have changed how often you get to sit in front of a client at all. Start with the mechanism that changes your economics, then let the design tools be the visible, satisfying reward once the boring machinery is already paying for itself.
The AI for Architects Self-Assessment Scorecard
Talk is cheap. Before you spend a dollar or an hour, you need an honest read on where your firm actually stands. Score each question from 0 to 3, then total it. Be brutal. Aspiration is not a score.
| # | Question | 0 = Never / No | 1 = Rarely | 2 = Sometimes | 3 = Systematically | |
|---|---|---|---|---|---|---|
| 1 | Do you use AI to draft or accelerate proposals and RFP responses? | |||||
| 2 | Do you track your bid win rate and proposal turnaround time as numbers? | |||||
| 3 | Is your marketing and lead nurture automated rather than ad hoc? | |||||
| 4 | Do you use AI to generate or cross-check specifications against drawings? | |||||
| 5 | Do you produce early design or massing options with generative tools? | |||||
| 6 | Is repetitive documentation and back-office admin partly automated? | |||||
| 7 | Do you forecast staffing and utilization across your project pipeline? | |||||
| 8 | Does at least one leader own AI adoption as an explicit responsibility? | |||||
| 9 | Have you redesigned any workflow around AI, not just added a tool to it? | |||||
| 10 | Do you measure the financial return of any AI initiative you have run? |
Add up your score out of 30, then find your band below.
| Total Score | Band | What It Means | Your Next Move | |
|---|---|---|---|---|
| 0 to 7 | Standing Still | You are in the majority. Fully exposed to faster competitors. | Pick one workflow. Start the 30-day plan below. | |
| 8 to 15 | Dabbling | Individuals experiment, but nothing is systematized. | Move from adoption to transformation. Redesign one process end to end. | |
| 16 to 23 | Building | Real systems exist in at least one function. | Scale what works, add measurement, and stop the pilots that do not pay. | |
| 24 to 30 | Compounding | AI is changing your unit economics. | Defend the lead. Push into capacity and utilization, the hard, high-value frontier. |
Most firms reading this will land between 0 and 15. That is not an insult. It is your starting line, and it is a far better place to start than where your competitors think they are.
A Practical 30, 60, 90 Day Roadmap
Strategy without sequencing is a wish. Here is the exact order I would run this in, because sequencing is where most transformations die. You do not boil the ocean. You win one thing, prove the number, and use that credibility to fund the next.
| Phase | Focus | Concrete Actions | Success Metric | |
|---|---|---|---|---|
| Days 1 to 30 | One workflow, one win | Pick your most painful repetitive process (usually proposals). Map it. Rebuild it around AI. Assign one owner. | First AI-assisted proposal shipped; hours-per-proposal measured before and after. | |
| Days 31 to 60 | Prove and measure | Run the new workflow on real projects. Track turnaround, win rate, and hours saved. Kill anything that does not pay. | A clean before-and-after number you can show partners. | |
| Days 61 to 90 | Expand and systematize | Add a second lever (marketing automation or documentation). Write the standard operating procedure. Train the team on the redesigned process, not the tool. | Two workflows running as systems; a documented playbook others can follow. |
Notice what is missing from this roadmap. There is no "evaluate fifteen vendors" phase. There is no six-month strategy committee. You choose one painful, expensive, repetitive workflow, and you rebuild it. Everything else follows from proving that first number. I have run this exact sequence, one win at a time, in every business I have transformed, and the discipline of it is the whole point.
Why You Start With the Boring Workflow
Founders and principals have an instinct to start AI with the exciting thing, usually design or rendering, because it demos well. That instinct is wrong, and it is why so many firms stall.
You start with proposals or documentation because those are measurable, repetitive, and directly tied to money. You can prove the return in weeks, not quarters. That proof buys you the internal credibility and the reclaimed hours to fund the harder, higher-value work later. Start with the flashy stuff and you will have a nice render and no business case. Start with the boring stuff and you will have a number that funds everything else.
The Investment Question: What This Actually Costs
Principals always want to know the price before the plan, which is backwards, but let me answer it honestly because vague answers help no one. Here is a realistic tiering of what getting serious about AI for an architecture firm costs, and what you should expect from each level.
| Tier | Typical Monthly Investment | What You Get | Best For | |
|---|---|---|---|---|
| Starter | Low, a few subscriptions | Off-the-shelf AI tools, individual productivity, one automated workflow | Solo practices and small firms testing the water | |
| Operational | Moderate | Integrated tools plus a part-time internal owner redesigning one or two workflows | Firms of 5 to 30 ready to systematize | |
| Transformational | Significant, a real line item | Custom workflows, integration with your BIM and CRM, dedicated ownership, measured return | Firms of 30-plus scaling AI across functions |
The honest truth is that the cost of the software is almost never the real cost. The real investment is attention, ownership, and the discipline to redesign a process rather than paper over it. A firm that spends little on tools but assigns a sharp owner to genuinely rebuild its proposal workflow will crush a firm that buys expensive software and changes nothing.
Build, Buy, or Bring In Help
There is a recurring decision here: do you hire someone in-house to own AI, buy tools and figure it out yourself, or bring in outside help to architect the transformation. There is no universal answer, but there is a framework for deciding, which I lay out in detail in my analysis of AI consulting versus hiring in-house. The short version: for a first, well-scoped transformation, external help that has already made these mistakes elsewhere is usually the fastest path to a proven number. For ongoing operation, ownership belongs inside your firm.
If you are weighing whether any of this pays off at all, the discipline of measuring return is non-negotiable, and I walk through exactly how to calculate it in my AI ROI for business guide. A firm that cannot state the return of its AI initiatives in a single sentence is not doing transformation. It is doing theater.
This is also the natural moment to bring in a second set of eyes. Booking a focused strategy session, with someone who has actually moved these numbers in real businesses, will save you months of wandering through vendor demos and dead-end pilots. The goal of that session is not to sell you software. It is to identify the one workflow whose redesign will pay for the entire effort.
What AI Will Not Do for Your Firm
I am a founder, and founders who oversell get found out fast. So let me be equally clear about the limits, because the hype in this space is dangerous and the AIA research shows that the profession is right to be cautious.
AI will not design your buildings. It will not replace judgment, taste, or the relationship that wins a commission. Nearly all architects surveyed by the AIA expressed concern about the accuracy and reliability of AI outputs, and they are correct to. Anything AI produces in your firm, a spec, a fee narrative, a drawing check, needs a licensed professional's review. The technology is an accelerator of expertise, never a substitute for it.
The Human in the Loop Is Not Optional
Every mechanism I described earlier has a human at the decision point. The AI drafts the proposal; the partner sharpens the strategy and owns the number. The AI flags the spec conflict; the architect makes the call. The AI predicts the staffing gap; the principal decides the hire. Remove the human from those loops and you do not get efficiency. You get liability. The firms that win with AI are the ones that use it to make their best people faster, not to replace their people with something cheaper and dumber.
This is exactly why the "list of AI tools" approach fails. Tools do not know your firm, your clients, or your risk tolerance. A transformation designed around your actual economics, with your people at the decision points, is the only version of this that survives contact with a real project.
The Five Ways Architecture Firms Botch This
I have watched enough businesses attempt technology transformation to know that failure follows patterns. Here are the five I see most often when firms approach AI, and how to sidestep each. If you recognize your firm in any of these, that recognition alone is worth more than any tool you could buy.
One: Chasing Tools Instead of Fixing Workflows
The most common failure, and the one I keep hammering, is treating this as a shopping problem. A firm evaluates a dozen platforms, buys three, and wonders six months later why nothing changed. The tool was never the constraint. The broken, undocumented, senior-time-devouring workflow was. Fix the workflow first, then choose the smallest tool that supports it.
Two: Starting With the Exciting Thing
Principals gravitate to design and rendering because they demo beautifully in a partner meeting. But those wins are hard to measure and slow to pay. Firms that lead with the flashy application burn their initial enthusiasm on something that never produces a defensible number, and the momentum dies. Lead with proposals or documentation, prove the return, and buy yourself the credibility to do the exciting work later.
Three: No Owner, No Outcome
When AI is everybody's job, it is nobody's job. The single strongest predictor of whether a firm actually transforms is whether one named person owns it as an explicit responsibility, not a hobby they squeeze in between projects. This does not require a new hire. It requires a decision. Name the owner, give them the mandate, and hold them to a number.
Four: Skipping Measurement
A firm that runs an AI initiative without a clean before-and-after number is flying blind and, worse, cannot defend the investment when a skeptical partner asks. Measurement is not bureaucracy. It is the mechanism that lets you kill what does not work and double down on what does. Hours per proposal, win rate, utilization. Pick the metric before you start, not after.
Five: Trying to Do Everything at Once
The overreach failure kills more transformations than caution ever will. A firm gets excited, launches five initiatives, staffs none of them properly, and watches all five stall. The discipline of one workflow, one owner, one proven number, then expand, is not timidity. It is how compounding actually works in the real world. My general operating model for sequencing this correctly, independent of industry, sits in the practical framework I reference throughout this piece, and the sequencing discipline is the part most firms skip.
How This Fits the Broader AI Shift
Architecture is not special in its exposure to AI, and that is good news. It means you can learn from every other professional services business that has already walked this path. The mechanisms that work for a law firm, an engineering consultancy, or a marketing agency work for you, because the underlying economics, selling expertise and time, are the same.
That is why the strongest thing an architecture principal can do is stop thinking of this as "AI in architecture" and start thinking of it as running a smarter professional services business that happens to design buildings. The frameworks are portable. If you want the general operating model, independent of industry, my AI implementation business practical framework is the backbone I return to for every business I run, and it applies to your firm without modification.
The broader picture is unambiguous. Research houses from Gartner to the analysts at Harvard Business Review all point to the same conclusion: AI is moving from experimental to operational across every knowledge industry, and the gap between leaders and laggards is widening, not closing. Architecture is simply later to the curve than most, which is precisely why the window is still open for you.
Making It Real: A Sequence, Not a Leap
Let me close the practical loop. If you read this and feel the pull to go transform everything at once, resist it. Transformation at a firm dies from overreach far more often than from caution.
Here is the discipline. Run the self-assessment above and get your honest number. Pick the single most painful, repetitive, expensive workflow, almost certainly proposals or documentation. Give it one owner. Rebuild it around AI in thirty days. Measure the before and after. Then, and only then, expand to a second lever. One win funds the next. That is how compounding works, and compounding, not any single tool, is what eventually reprices your entire firm's economics.
If you want a shortcut through the wandering phase, this is where a strategy session earns its keep. Sitting down with someone who has actually delivered these numbers, a 20 percent capacity lift, a revenue jump from better predictive decisions, a 30 percent sales increase from AI marketing, means you skip the expensive lessons and go straight to the sequence that works for a firm your size. You do not need to be the one who makes every mistake first. You need to be the one who moves before your competitors do.
The Bottom Line for Architecture Principals
The data is not ambiguous. Most architecture firms have not implemented AI in any meaningful way, the sector is slow, and the firms moving now are large and getting further ahead. That is not a threat. For a sharp principal, it is the clearest arbitrage opportunity your profession has offered in a generation.
AI for architects is not about drawing prettier renders. It is about winning more bids, acquiring more clients, delivering with fewer wasted hours, and squeezing more capacity out of the team you already have. Those are business outcomes, and they are exactly the outcomes I have delivered, with the same mechanisms, in hospitality, sport, healthcare adjacent services, and marketing. The costume changes. The engine does not.
Pick one workflow. Prove one number. Then compound. Your competitors are standing still. That is the whole opportunity, and it will not stay open forever.
Frequently Asked Questions
Is AI going to replace architects?
No. AI will not design your buildings or replace the judgment, taste, and relationships that win commissions. What it will do is make your best people dramatically faster at the repetitive, expensive work around the design: proposals, documentation, coordination, marketing, and planning. The architects at risk are not the ones who use AI. They are the ones whose competitors use it and they do not.
What is the single best place for an architecture firm to start with AI?
Proposals and RFP responses, almost always. They are repetitive, expensive in senior time, directly tied to revenue, and easy to measure. AEC firms spend on average more than two weeks on a single RFP, and firms using AI report cutting that by a third or more. You can prove a return in weeks, which funds everything else.
How much does it cost to get serious about AI for architects?
The software is rarely the real cost. Off-the-shelf tools start cheap. The real investment is attention and ownership: assigning someone to genuinely redesign a workflow rather than paper over it. A firm that spends little but redesigns its proposal process well will beat a firm that buys expensive software and changes nothing.
We are a small firm. Is this only for large practices?
The opposite. Large firms lead adoption today, but small firms can move faster because they have fewer layers and less process inertia. A five-person practice can rebuild its proposal workflow in a month. A two-hundred-person firm needs a committee to agree on lunch. Your size is an advantage if you use it.
Will AI outputs be accurate enough to trust on real projects?
Not without review, and you should never treat them as if they are. Nearly all architects surveyed by the AIA are concerned about AI accuracy, correctly. Every AI output in your firm, a spec, a drawing check, a fee narrative, needs a licensed professional at the decision point. AI accelerates expertise. It does not replace it.
How is this different from just buying AI software?
Buying software is adoption. Rebuilding a workflow around it is transformation. McKinsey's research shows the firms getting real financial return redesigned their processes, while the majority that just bought tools are stuck in pilot purgatory. The tool is downstream of the workflow decision, and it barely matters by comparison.
Should we hire someone in-house or bring in outside help?
For a first, well-scoped transformation, outside help that has already made these mistakes elsewhere is usually the fastest path to a proven number. For ongoing operation, ownership belongs inside your firm. Many firms do both: bring in help to architect the first win, then keep the running of it internal.
How do I know if any of this is actually paying off?
You measure it, in one sentence, per initiative. Hours saved per proposal. Change in win rate. Change in utilization. If you cannot state the return of an AI initiative in a single number, you are not doing transformation, you are doing theater. Measuring return is the discipline that separates firms that compound from firms that dabble.
How long before we see a real return?
Weeks, not quarters, if you scope it correctly. That is the entire reason I insist you start with proposals or documentation rather than design. A proposal workflow rebuilt in thirty days produces a measurable before-and-after number, hours saved and turnaround cut, inside the first month of live use. The mistake that pushes returns out to a year is trying to transform everything at once. Narrow the scope and the payback window collapses.
Does firm size or specialty change the approach?
The sequence stays the same regardless of whether you do residential, commercial, healthcare, or civic work, because the underlying economics of selling expertise and time are identical across specialties. What changes is which workflow is most painful. A firm drowning in public-sector RFPs starts with proposals. A firm buried in complex technical documentation starts there. Run the self-assessment, find your most expensive repetitive process, and start with that one. The method is portable. The starting point is personal to your firm.
What should I do first, this week?
Run the ten-question self-assessment in this article and get your honest score. Then pick the single most painful, repetitive, expensive workflow in your firm and commit to rebuilding it around AI in the next thirty days, with one named owner. If you want to skip the wandering phase, book a focused strategy session with someone who has actually moved these numbers, and use it to identify the one workflow whose redesign pays for the entire effort. You do not need to make every mistake yourself. You need to move before your competitors do.
