Product work · AI product design

Prompt Studio: from developer tool to product

Redesigning NetSuite’s prompt customization tool in the earliest days of generative AI — when nobody, anywhere, knew yet what “good” looked like.

AIResearch

Prompt Studio is a publicly released feature; screens shown with sample data.

Role
Lead designer, partnering with researchers on usability testing
When
2024–2025, with discovery starting when generative AI was barely a year into enterprise software
Scope
Redesign of the prompt authoring & management experience for NetSuite administrators
Outcome
Shipped; restructured from a developer form into a task-ordered product

The problem: AI administration nobody had designed before

Prompt Studio lets administrators customize the AI prompts behind NetSuite’s features — tuning how AI writes for their company’s records, fields, and tone. When this project started, generative AI was brand new to enterprise software. There were no conventions to follow: prompt authoring tools barely existed anywhere, let alone patterns for making them usable by admins rather than engineers.

Before: a developer’s form

The original version exposed the system’s internals directly: raw template syntax typed by hand, model parameters like Top K and frequency penalty presented flat on one page, preview buried at the bottom. It worked — for the people who built it. For an administrator trying to make AI write purchase descriptions properly, every step was a guess.

Before the revamp: the original Prompt page as one flat developer form — prompt settings in a row, a raw template field containing hand-typed handler syntax, model settings like Cohere Command R, temperature, Top K, Top P, and frequency penalty exposed flat, and the preview buried at the bottom of the page.

Before: one flat form — raw template syntax typed by hand, model parameters in a row, preview at the bottom. Built by developers, for developers.

Evidence first: watch people fail, study the field

Rather than debating opinions about a tool with no precedent, we gathered evidence two ways: usability testing with our researchers to see exactly where people fell off, and competitor research across the emerging landscape of prompt tools to learn what conventions were starting to form. The sessions were humbling in the useful way — the flat form’s order matched the system’s schema, not the user’s thinking.

After: structured around the task, not the schema

The redesign reordered everything around how an admin actually thinks:

After the revamp: the New Prompt page restructured around the admin's task — a left rail ordered as Select Target (prompt type, record type, field name), then Basic Prompt Settings in plain language, then Model Settings grouped at the end; the main area holds the template with guided variable insertion (pick a variable type, pick a variable, add to template), a variables-in-template table where sample values can be entered, and a preview pane that shows the prompt and its generated result side by side.

After: the page follows the admin’s task order — select the target, then settings in plain language; variables are inserted from dropdowns instead of typed syntax, and the preview runs the prompt against sample values before anything is deployed.

The structure was deliberately anchored to the user’s task, not to any particular UI kit — the kind of foundation that outlives platform changes, because “what am I trying to do first?” doesn’t change when the technology underneath does.

Outcome

What I learned

When a technology is too new to have conventions, the user’s task order is the only stable foundation — schemas change, models change, but “what am I trying to do first?” doesn’t. Test early precisely because nobody knows what good looks like yet: in a space with no conventions, a handful of usability sessions makes you the best-informed person in the room.