← All posts

How The Dojo gives every workflow creative context.

CompanyThe Dojo
CategoryCreative performance agency
SurfaceMCP · Agentic workflows
The Dojo×Upspring
What I like about Upspring is that it doesn't try to replace how we think or work. It gives our AI the creative context it was missing, and lets us use that inside the workflows we already have.
Arsh ArshFounder · The Dojo

About The Dojo

The Dojo is a creative performance agency that builds and scales emerging consumer brands. Their work runs the full distance, from the strategic read on a category to the script that goes into production this week.

They had already moved fast on AI. Agentic workflows for research and planning, Claude in daily use across the team, a way of working the agency actually believed in. What they were missing was not intelligence. It was something for that intelligence to look at.

The challenge

Smart AI, no view of the market.

An agency's edge is knowing what is working in a category before the client does. That knowledge lives inside live ads: the hooks running right now, the claims spreading, the advertisers quietly scaling. None of it is readable by a model on its own.

Creative isn't data

Video, images, copy, and landing pages can't be searched or compared at scale.

Manual research doesn't scale

One brand's competitive scan is an afternoon. A roster of them is a department.

Context stayed in one head

Whoever did the research had the picture. Nobody else's AI did.

Building the collection and creative processing themselves was never realistic, and adding another dashboard to log into would have meant changing how the team already worked.

Why Upspring

Connected once, available everywhere.

The Dojo connected Upspring over MCP. That is the whole integration. No new interface, no export step, no separate research tool sitting beside the work.

It means the creative context shows up in two places at once. Their agentic workflows can pull live advertisers, ads, creative attributes, and scale grades mid-run, without a person fetching anything. And every employee's chat has the same access, so a strategist can ask about a category in plain language and get an answer grounded in what is actually live.

The same data underneath both. One connection, not one per person and one per pipeline.

In practice

The same context for strategy and for scripts.

This is what makes it stick. The high-level work and the production work draw on one source.

On the strategy side, the questions are about the market. Who is scaling in this category, what claims are they leading with, where is the whitespace, what changed this month. Advertisers arrive graded on spend, traffic, and sales, so a serious operator is easy to tell from a test budget.

On the practical side, the questions are about the next asset. What hooks are working in this format right now, which offers are showing up in the top spenders, what has this competitor shipped since we last looked. The writer asks in the same chat they were already working in, and the answer is specific enough to build a script on.

"We can build on that and then layer in anything," Arsh says. It has been a key driver in helping them grow a set of fast-growing emerging brands, ULTRA among them.

The result

AI that fits the way they already work.

The Dojo did not change their process to adopt Upspring. They kept their workflows, their tools, and their judgment, and gave all of it something to reason with.

0pipelines to maintain
1connection, every workflow and every chat
10×fewer tokens vs. raw creative

The thinking is still theirs. What changed is that it now happens with the market in view.

Build on the same data layer.