The Question You Cannot Answer Fast Enough
A developer sends you an address. They want to know if the site is worth pursuing. What can it hold, what constrains it, what hides beneath the surface that could sink the pro forma six months from now. You know how to find the answers. Zoning, dwelling unit limits, utility capacity, transportation access, environmental flags. You also know what it costs you to find them. Days of digging through county records and planning portals, chasing utility providers, cross-referencing maps that do not agree with each other. By the time you have a real answer, the developer has often moved on, or the opportunity has cooled.
This is the quiet tax on early-phase work, and AI site feasibility analysis is beginning to erase it. That is the thread I kept pulling on in my conversation with Gayathri Venkataraman, known to most people as G3, founder and CEO of GeoProspex. She spent more than two decades leading AI product development at Microsoft and DocuSign before turning that experience toward one of the most underserved moments in our entire process: the point before design begins, when someone has to decide whether a piece of land is worth another dollar of anyone’s time.
What struck me was not the technology alone. It was what the technology means for firms like yours. The work that used to take roughly 120 hours, G3 told me, can now happen in minutes. Sit with that number for a moment, because the number is not the point. The point is what you do with the weeks you get back.
Why Speed Is Not the Real Story
It would be easy to hear “minutes instead of weeks” and file it under efficiency. Faster is nice. Faster is not transformation.
The transformation is in what the speed unlocks for how you position your firm. G3 described a pattern already showing up among her clients. When feasibility findings take minutes instead of weeks, a firm can hand a developer early insight as a low-cost or no-cost first step, and use that generosity to earn the larger design contract that follows. The tool does not replace your judgment. It lowers the cost of demonstrating your judgment early enough to matter, when the client is still deciding who to trust with the whole project.
That is a positioning move, not a software feature. You are no longer the firm that says “give me two weeks and a retainer before I can tell you anything.” You are the firm that walks into the room already knowing what the site can hold and what stands in the way. Feasibility analysis stops being invisible back-office labor and becomes the thing that wins you the work.
There is a deeper reframe here worth naming plainly. Feasibility has always been treated as a cost to absorb. Reframed through AI, it becomes a business development tool. The same hours that used to drain your margin can now open the door.
The Accuracy Problem Every Architect Should Understand
I pressed G3 on the concern I hear constantly from architects experimenting with AI, and the concern I feel myself. The large language models are confident. They are not always correct. Ask a general AI tool about zoning or utility capacity and it will hand you an answer in fluent, authoritative prose, and you will have no way of knowing whether it is true. Ask it the same question twice and you may get two different answers. In feasibility work, a confident wrong answer is worse than no answer, because you might build a recommendation on top of it.
Her response clarified something important about how to evaluate any AI tool that enters your workflow. GeoProspex pulls from authoritative data sources rather than scraping the open internet, and when the authoritative source does not have the answer, the tool says it does not know rather than inventing something plausible. It also holds consistency, so the same question returns the same result every time. When I noted that a general chatbot never tells you it does not know, G3 agreed, and that gap is the whole game. A tool that admits its limits protects you. A tool that always sounds certain exposes you.
This is the standard you should carry into every AI decision you make for your firm. Not “is it impressive,” but “does it know the difference between what it knows and what it is guessing.” Your license, your reputation, and your client’s capital are downstream of that difference.
Small Firms Are Built to Win This Moment
Here is where the conversation turned to the part I most want you to hear. There is a persistent belief that emerging technology favors the large firm, the one with a technology budget and a dedicated innovation team. G3 sees it the other way, and so do I.
Small firms have the greatest opportunity in this shift, and the reasons are structural. You can make a decision to try a new tool this afternoon. You do not need a committee, a procurement cycle, or three layers of approval. You can adopt something on Monday, learn it by Wednesday, and use it on a live project by Friday. That speed of decision is a genuine advantage, and it is one large firms cannot buy. The tools now reaching the market let a solo practitioner or a team of four take on analysis that used to require resources only a big firm could marshal. For the first time, the size that once limited you can become the thing that lets you move faster than everyone around you.
G3 put the mandate simply when I asked what a small firm architect can do today to build a better business for tomorrow. Embrace AI. Disrupt yourself before the change disrupts you. Go after it proactively rather than waiting for it to arrive like a storm you did not prepare for. Try new tools. Take charge of how you craft your future rather than letting the future be handed to you.
The Fear Is Real, and It Is Not the Enemy
I want to be honest about the thing sitting underneath all of this, because pretending it does not exist helps no one. There is real fear in our profession right now. Fear of handing sensitive project information to a company you just met. Fear that the tool will replace the judgment you spent a career building. Fear of paying for something when a free chatbot seems close enough. G3 named all of it directly, and she did not dismiss any of it.
The fear is not a flaw in you. It is information. It tells you to ask hard questions about data privacy, about accuracy, about whether a vendor will still exist in two years. Those are the right questions. G3 was candid that our industry moves slowly and that trust is earned over time, not claimed in a demo. She has watched clients grow comfortable only after working alongside the tool a few times, watching the fear recede as understanding grows. That is the honest arc of adoption. Excitement, then fear, then gradual trust as the technology becomes an ordinary part of the workday.
The mistake is not feeling the fear. The mistake is letting it make your decisions for you while the firms willing to learn move ahead.
Do Not Lose the Thing That Makes You Valuable
For all her conviction about AI, the moment from our conversation that stayed with me was about what AI cannot do. G3 talked about humans in the loop, and then corrected the usual framing. We tend to say humans are needed for validation, to check the machine’s work. She argued the deeper reason humans belong in the loop is empathy.
A tool can consolidate zoning data. It cannot sit across from a client and understand why this project matters to them, what they are afraid of, what they are really trying to build in their life and not just on the parcel. That understanding is the source of your creativity and the reason clients hire a person instead of a platform. I shared with G3 that a longtime member of our community recently wrote to me, concerned we were leaning on AI too heavily in our own work. I was grateful for it. It was a reminder that EntreArchitect is about people helping people, and that the tool should sharpen what makes us human rather than sand it away.
This is the balance every small firm entrepreneur architect has to strike now. Use AI to erase the hours that were never the point. Protect the judgment, the empathy, and the human connection that were always the point. If you become the architect who runs every decision through a machine, you will slowly stop being the architect anyone needed in the first place. Adopt the speed. Keep the soul.
Where This Leaves You
The firms that will thrive in the next few years are not the ones with the biggest technology budgets. They are the ones intentional enough to adopt the right tools quickly and grounded enough to keep their humanity while they do it. AI site feasibility analysis is one clear, practical example of that principle in action, a place where the busywork disappears and your real value moves to the front of the engagement. Start there, or start somewhere else. What matters is that you start, with clear eyes about both the opportunity and the responsibility.
To hear the full conversation with G3, including how GeoProspex handles community sentiment for large projects and where the technology is heading next, listen to the full episode at https://entrearchitect.com/671.
And if you want to build a better, more resilient firm alongside a community of small firm architects navigating these same questions, I invite you to join us inside EntreArchitect Network at https://entrearchitect.com.
Leave a Reply