AI workflows for 3D and archviz studios

These are the four places where an AI assistant saves a small 3D studio real time, written as the prompts we use, with what the assistant does in the project tool at each step and where it stops.

By · Updated · 6 min read

On this page
  1. The short answer
  2. Before you start: two habits
  3. Workflow 1: brief to task list
  4. Workflow 2: feedback email to tickets
  5. Workflow 3: the weekly client update
  6. Workflow 4: a first quote draft
  7. What each workflow uses
  8. Things to watch

The short answer

The assistant is useful in a 3D studio wherever a long, messy piece of text has to become a structured list: a brief, a feedback email, a week of notes, a scope for a quote. Connected to your project tool, it writes that list straight into the project. The four workflows below cover most of the admin on a typical archviz or product render job.

They assume Claude or ChatGPT connected to OIO through MCP. If you have not set that up, project planning with AI has the steps. They also work without a connection: you copy the result by hand, which is slower but still faster than writing it yourself.

Before you start: two habits

First, create the project in OIO yourself, with the client and the deadline. The AI connection cannot create projects or clients, and we prefer it that way: the job exists because you accepted it, not because a model read an email.

Second, tell the assistant the project by name at the start of every session, and ask it to list the existing tickets before it adds any. The create action in OIO reminds the assistant to check for duplicates, but a sentence from you makes it reliable.

Workflow 1: brief to task list

An archviz brief usually arrives as a PDF, a set of drawings and an email that changes half of it. The deliverables are scattered across all three.

What to paste or attach: the brief and the email. Then ask:

Read the OIO project "Harbour Lofts, phase 2". List its tickets first.
From this brief, make one ticket per deliverable (each still image,
each animation), named the way the client names the view.
Due dates: first draft on the dates in the brief, or 5 working days
before the final deadline where none is given.
On each ticket, add to-dos for: camera approval, modelling, materials,
lighting, first draft, round 1 changes, round 2 changes, final.
Add the brief's lines about that view to the ticket's note.
Then list what is missing from the brief.

What happens in OIO: the assistant reads the project, creates a ticket per view with its due date, adds the to-dos one by one, and appends the relevant part of the brief to each ticket's note with today's date. The tickets show up on the kanban and the timeline straight away, so overlapping jobs are visible before you confirm the schedule.

Where it stops: it cannot assign tickets to people or attach the drawings. You do that in the project, and freelancers get their ticket link from there. The list of missing information stays in the chat, which is the right place for it: it becomes your reply to the client.

Name tickets as you want the client to read them. The status link shows ticket names and their status, so "Kitchen, eye level, dusk" is better than "KIT_v2_fix".

Workflow 2: feedback email to tickets

Feedback on renders arrives as a long email, often with notes from three people on the client side, sometimes contradicting each other. Turning it into changes per view is the job nobody wants. How to run the rounds themselves is in archviz client feedback rounds; this is the typing part.

Here is the client's feedback on round 1 for "Harbour Lofts, phase 2".
For each view mentioned, find its ticket, add every change as a
separate to-do starting with "R2:", and append the client's original
wording to the ticket's note. If a note does not clearly belong to a
view, do not guess: list it for me. List any notes that contradict
each other.

What happens in OIO: for each view, the assistant finds the ticket, adds the changes as to-dos and appends the original wording to the note, so the exact words are on record next to the work. A view with no ticket yet can get a new one if you allow it. The contradictions come back in the chat for you to put to the client.

Where it stops: it cannot see the images. If the feedback says "the sofa on the left", the assistant does not know which sofa that is. Keep a drawn markup from the client next to the job, and let the assistant handle the text.

Workflow 3: the weekly client update

Clients ask "where are we?" more often than they give feedback. A short weekly update stops most of that.

For each of my active projects, read the tickets and tell me what was
finished this week, what is in progress and what is waiting on the
client. Then draft one status line per project for the client, under
300 characters, no internal notes, no prices. Show me the lines before
posting anything.

After you approve them: "Post those as status updates, and give me each project's status link."

What happens in OIO: the assistant reads the projects and tickets, drafts the lines, and after your go-ahead posts each one as a status update on the project. Status updates appear on the client's password-protected status link, next to the deliverables and their statuses. It then fetches the existing status link for each project so you can paste it into an email.

Where it stops: it cannot create a status link if the client does not have one yet; you make that in OIO. It cannot send the email either, unless your assistant has its own email connection, and we would still read it first. Status lines are visible to the client, which is why the prompt asks to see them before they go out.

Workflow 4: a first quote draft

Pricing is your decision, but the line items are tedious. An assistant reading a brief is good at listing everything that will take time, including the things people forget to quote: camera approval rounds, extra rounds, the second aspect ratio for social.

From this brief, list every deliverable and every task that will take
time, as quote lines with a unit (per image, per second of animation,
per round, per day). Leave prices empty. Then list what the brief does
not say that would change the price.

What happens in OIO: nothing yet, on purpose. The OIO connection has no access to quotes, invoices or money. You put your prices on the lines and make the quote in OIO from the project, or with the free quote template if you are not on OIO. If you need a number to start from, the day rate calculator and how to price archviz images cover that.

What each workflow uses

WorkflowReadsWritesClient sees
Brief to task listProjects, ticketsTickets, to-dos, notesTicket names and status
Feedback to ticketsTicketsTo-dos, notes, sometimes ticketsNothing new
Weekly updateProjects, tickets, status linkStatus linesThe status lines
Quote draftNothing in OIONothing in OIONothing

Things to watch

  • Notes from clients and freelancers are text the assistant reads. If one of them contains something that looks like an instruction, OIO marks it as untrusted for the assistant, but read what the assistant says it changed.
  • Every change made through the connection is in Logs with the connection's name, and every one can be reversed in the app.
  • The connection only writes to projects you own. On a project someone shared with you, it can read but not change.
  • Whatever the assistant reads goes to its provider. Keep that in mind on jobs under NDA, and use a read-only connection or none.

For how small 3D teams run whole jobs in OIO, see OIO for 3D studios and OIO for archviz studios. Plans with the AI connection start at Pro, on pricing.

Questions

Can ChatGPT or Claude manage an archviz project?

They can turn briefs and feedback into tickets and to-dos and draft client updates. Connected to OIO they write those straight into the project. They cannot see renders, assign work, or make quotes and invoices through the OIO connection.

Can the AI read my render feedback images?

Not through the OIO connection, which handles text only: tickets, to-dos, notes and status lines. Many assistants can look at an image you attach to the chat, but they still cannot tell which object a client means as reliably as you can.

Will the client see what the AI writes?

Only ticket names, ticket statuses and status updates show on the client's status link. Ticket notes and to-dos stay internal. Ask the assistant to show you status lines before it posts them.

OIO is project management for small 3D and archviz studios: a project per client job, tickets on a kanban and a timeline, a link for each freelancer and one for the client. On Pro and up, Claude and other MCP apps can read and update it. Create a free account.

More guides