Making an agentic content tool usable for non-marketers

Robynn's agents could plan a month of content, but only technical marketers could use them

Robynn AI is an agentic marketing platform. Its Content Studio takes what a company wants to write about, proposes a content plan, and has agents draft the pieces. In practice, the flow was long, the language assumed a marketing background, and it was hard to tell what the agent was doing.

My role was to redesign it so founders and generalists could use it too

Researching how marketers work

I found that marketers create both content plans and single pieces in the same six steps

Key design decisions

Problem

Each screen had too many buttons and tasks competing for attention

Solution

I cut the number of choices on each screen and used the accent color to highlight the most common tasks

Choice overload: "The tendency for people to get overwhelmed when they are presented with a large number of options."

Problem

The interface relied on marketing jargon, which created friction for non-technical users

Solution

I rewrote the interface in plain language, so no one had to look up a term

Match between the system and the real world: "speak the user’s language and avoid internal jargon"

Problem

The writer copilot was a powerful editing tool, but users didn't know how to use it

Solution

I redesigned the copilot to work like the AI chat tools people already use

Jakob's Law: "Users spend most of their time on other sites. This means that users prefer your site to work the same way as all the other sites they already know."

Problem

While the agent worked, users only saw "Generating research…", so they couldn't tell what it was doing or how long it would take

Solution

I added step-by-step progress indicators so users could see what the agent was working on and how close it was to done

Visibility of system status: "keep users informed with timely feedback."

I handed off a design.md file that engineers and AI tools build from

The rules mattered more than the tokens. Tokens tell a builder what exists. Rules tell them what to choose, which is where inconsistent screens come from. Every wireframe element was labeled with its component, so styling became a lookup. Rebuilding every other feature on a new system would have cost too much, so my redesign shipped on the existing one.

Users reported that Content Studio made it easier to accomplish their goals

Walk through the key flows, from planning a month of content to reviewing a finished piece.

Reflections

Working as the solo designer on a small team meant I had to make, and defend, a lot of design decisions. It also meant taking ownership of and accountability for them. For instance, I knew that cutting the research report generation would be a tradeoff, but it was necessary to make the feature more user-friendly.

AI tools helped me work within tight time constraints. With more time and budget, I'd have run usability tests with founders and generalists before handoff.