Ralph: The AI Project Assistant We Built Because We Got Tired of Chasing Our Own Updates

Every studio hits the same wall eventually. The work is going fine, but the admin around the work, chasing statuses, writing update emails, remembering who asked what three weeks ago, starts eating hours that should go into the actual 3D, motion or web build. We hit that wall clearly enough that we built our own answer to it: Ralph, an AI project assistant wired directly into the project management tools we already run every job through.

Ralph is not a chatbot bolted on top of our workflow. It sits inside the same system our producers use to track tasks, deadlines and client threads, which means it has the same context a human project manager would have rather than a generic assistant guessing from a prompt. It can pull together where a project actually stands, draft the update before someone has to write it from scratch, and flag the tasks that are quietly drifting past their date before a client has to ask.

The instinct to build this ourselves, rather than reach for an off the shelf tool, came from the same place as most of our internal builds: the generic version does most of what you need and gets in the way on the rest. Client project management in a studio doing 3D, motion, web and branding work in parallel does not look like project management in a software team. A single job might have a render pending sign off, a copy deck waiting on legal, and a dev sprint running underneath it, all reporting to the same client at the same time. Off the shelf tools want to flatten that into one kind of task. Ralph is built around the fact that it is not one kind of task.

We were careful about where to draw the line. Ralph drafts, it does not decide, and nothing it writes goes to a client without someone on the team reading it first. The point was never to remove a person from the loop, it was to remove the blank page every producer used to start from when they sat down to write a status update on a Friday afternoon. That is a small thing until you multiply it across every live project every week, at which point it stops being small.

Building Ralph also changed how we think about the tools we sell to clients. A fair amount of our web and platform work now involves exactly this question, how much of a workflow should an AI agent handle directly versus draft for a human to approve, and having lived with that decision inside our own studio first means we are not guessing when a client asks us to make the same call for their team.