Services

Four ways in.

We work with design teams, front-end engineers and product teams bringing AI into what they build and how they build it. Every offer runs on a tool anyone can install, so you can check how we work before you ever talk to us. Timeframes below are typical. Scope and price come out of the first conversation.

Start a conversation

Drift Audit

Runs on Stage1 · About two weeks

See where your product and your design system have drifted apart. We capture the live product, measure it against your Figma library, your token files and your source, and trace each difference to its cause. Then we walk your team through what to fix first.

It’s the quickest way to find out whether we’re useful to you, and you keep the evidence either way.

What we ask for
  • The pages or journeys in scope. A standard audit covers up to 40 pages or five signed-in journeys.
  • A test account for signed-in areas, as a saved browser login. Never passwords in links.
  • Your exported .fig file, or access to the Figma library.
  • Your token files, if they live outside Figma.
  • Optionally, read access to the front-end repository, so a finding can point at its source.
  • Written permission to capture, unless the product is public and its robots rules allow it.
What you get
  • The Stage1 evidence export. Anyone on your team can open it in a browser, with nothing to install.
  • A written readout: what matches, each difference traced to its cause, what we couldn’t measure, and what to fix first.
  • A 60-minute walkthrough with your team.

Screenshots can contain customer data. We delete our copies 30 days after the walkthrough unless you ask us to keep them. The export is yours.

See a finished one: USWDS, measured against its own Figma kit

AI Workshops & Training

Runs on CMOS and TraceLab · Half a day to several weeks

For most teams bringing AI into how they build, the gap isn’t the model. It’s how people work, especially where the context lives and how decisions carry from one person or session to the next.

Formats we’d suggest, scoped together, remote or on site:

  • A working session with your team on where AI fits in how you build.
  • Hands-on workshops on your own project.
  • A longer program that sets up a shared way of working with AI, with context and decisions that carry between people and sessions.
  • Talks.

We run them on CMOS, which keeps missions, decisions and project context across agent sessions, and TraceLab, which answers questions from your research and shows the sources behind each answer.

Object-Oriented Design Systems

Runs on OODS Foundry · Several weeks

Keep the design system you have, and layer your product’s objects on it. We model those objects and the traits they share, connect them to your existing components and tokens, and set your agents up to build screens from them, with a receipt for every app they generate.

Product & UX Leadership

No tool needed · Ongoing, or for the length of a project

Complex product UX and information architecture, scoped accessibility reviews, and fractional design and systems leadership. Accessibility reviews cover the pages and states we agree on, and say what was tested and what wasn’t.

How an engagement starts

  1. A conversation. Thirty minutes with the people who’d do the work. Bring the product or workflow and what’s in the way. No pitch deck.

  2. A short proposal. Scope, timeframe, and exactly what we’d need from you: files, access, people.

  3. Start small. Most work starts with a Drift Audit or a single workshop, so both sides find out quickly whether it’s worth more.

  4. Keep going if it’s working.

Start a conversation

We reply within two business days.