Your practice finishes a project. The photographer comes, the images get filed, three go on Instagram, one ends up on the website six weeks later with a caption, and the rest sit in a folder nobody opens again.
Multiply that by every project of the last decade and you have the actual reason most practices struggle with architecture seo agency. Not a technical failure. Not a bad agency. A studio routine that produces excellent buildings and almost no usable record of them.
Why the website is rarely where the problem starts
Agencies get blamed for this, often unfairly. A practice hires someone, the work begins, and then it stalls because the agency is waiting for project details that nobody inside the practice has time to assemble.
The asymmetry is worth naming. Writing a project page takes an hour if you have the constraints, the planning history, the materials, and three decent photographs. It takes a fortnight if you have to chase a project architect who is mid deadline on something else, find out which planning conditions applied, and work out whether the client minds being named.
So the content never gets made, the retainer bills against technical work alone, and everyone concludes SEO does not work for architecture practices.
What should happen the week a project completes
This is the fix, and it is a process change rather than a marketing one.
Build a short handover into project completion, the same way you already handle final certificates and client sign off. Twenty minutes with whoever ran the project, capturing a fixed set of things: what the client wanted, what made the site difficult, what the planning position was, which materials were chosen and why, and what the practical outcome was.
That is the raw material for everything. A project page, a case study, a post, an award entry, a conversation with a journalist. Captured once, used repeatedly.
The practices that do this well treat it as part of closing a job rather than a marketing task, because marketing tasks get postponed and project closure does not.
The photography decision that quietly limits everything
Worth a separate mention because practices make this call early and live with it for years.
Commission photography that covers more than the hero shot. Detail images, context shots showing the street or the setting, and ideally something of the project in use. The hero image is for the homepage. The others are what make a project page usable, because a page with eight images showing how a building actually works is a far better answer to someone researching than one perfect elevation.
Also agree usage rights properly. A practice that cannot publish its own project photography without a conversation each time will publish less of it, which sounds obvious and happens constantly.
Who in the practice should own this
Almost never the person people assume.
A marketing hire is overkill for most practices and underpowered for the rest. What works better is a practice manager or an architect with genuine interest who owns the routine: collecting the handover notes, chasing the photography, approving what an agency drafts, keeping the business profile current.
Four or five hours a month, written into the role rather than assumed. That is usually the difference between an engagement that compounds and one that stalls in month three.
The failure mode is shared ownership, where the agency waits on the practice and the practice waits on the agency and a quarter passes with nothing shipped. Somebody has to be accountable, and in most practices that person already exists and simply has not been asked.
What this unlocks on the AI side
Here is the part that has changed recently and matters more than practices expect.
Systems like ChatGPT, Gemini, and Google AI Overviews answer questions about which practices to consider by reading structured text and third party mentions. They cannot see your buildings. They cannot assess your drawings. They work from what is written down and how clearly it is organised.
A practice with forty detailed project records, each describing a specific type of building in a specific place with its constraints and outcomes, gives those systems something substantial to work with. A practice with a gallery of beautiful unlabelled images gives them nothing at all.
This is why AI SEO services depend so heavily on what happens in the studio. The structural work, schema, entity clarity, content organised so a model can extract a clear answer, all of it needs source material. No amount of technical skill compensates for an empty archive. Expect roughly 60 to 90 days for initial movement in citation tracking once the material exists, and be wary of anyone guaranteeing placement inside a specific answer, since nobody controls those systems.
Test where you stand in two minutes. Ask one of those assistants to recommend architecture practices in your city and see whether you are named.
Catching up on a decade of unrecorded work
If you have ten years of projects with no usable record, do not attempt all of them. That decision is why most practices never start.
Pick the ten projects that best represent the work you want more of. Not your favourites, not the most awarded, the ones closest to the briefs you are chasing next year. Write those properly, one a fortnight.
The rest can wait indefinitely, and a few probably should, because a project you no longer want to be known for is not worth the effort of documenting well.
The test that tells you whether the routine is working
Six months from now, ask whoever owns this to tell you how many projects completed in that period and how many have a proper page.
If those two numbers match, the routine has taken hold and everything else becomes easier. If the second is zero, no agency is going to rescue the situation, because the thing they need from you is the thing that is not being produced.
Fix the studio habit first. The website follows it rather than the other way round.
