AI

AI.

Useful in the product. Useful in the work of making it.

Less brings design, product, frontend and AI capability to real tasks: the interaction people use, the workflow around it, and the system that helps a team build and maintain it.

See the work

Where it is showing up.

Platform processes

Less works with OpenAI on evolving their platform processes.

Ask about this work ↗

Merchant activation

Less explored an AI-assisted path through quotes, contracts and onboarding using product design and code prototypes.

See the prototype work ↗

Merchant activation prototype

Request the full case ↗

Design the work around the model.

An AI-generated offer is useful when a sales person can understand it, change it and move the merchant forward. For DoorDash, Less prototyped the product experience around a generated quote: commercial terms, human review, signature and activation in one working flow.

The work joins product flows, detailed interface design and code prototyping. It gives the team a working experience to review, rather than leaving the questions about AI behavior in a static screen.

DoorDash prototype screen showing an AI-generated merchant offer, proposed rates, contract access and a send-for-signature action — Prototype screen · AI-generated quote ready for review and signature

Turn generated output into a usable product.

The sales user sees the proposed plan and rates, opens the contract, adjusts the offer and decides when it is ready to send. The design connects AI output to the surrounding product: editable terms, an inspectable contract, a decision to send and a visible next state.

DoorDash prototype quote screen showing pending signature status and activity history — **After the send.** The quote, signature state and activity history stay visible, so the next action is clear.

DoorDash prototype deals screen showing merchant offers in draft, pending signature, signed and activated states — **Across merchants.** Deal states make it possible to track drafts, signatures and activation as operational work.

Selected screens from a working prototype. Want the decisions behind them?

Ask for a walkthrough

Where we can take responsibility.

AI product design

Work out what AI should do in the product, then design the interaction in enough detail to build and test.

Task flows, input and prompt interactions
Generated content, editing and source visibility
Review, approval, correction and undo
Waiting, streaming, uncertainty and failure states
Detailed screens and implementation review

Human–AI workflows

Design the whole task across people, AI and the systems a business uses.

Task, dependency and decision mapping
Human review and permission boundaries
Handoffs, exceptions and recovery paths
Internal interfaces and operational tools
Prototype testing with the people doing the work

Design systems for AI-assisted development

Refactor the library and the design-to-development workflow so tools can use the team's actual components and rules.

Component, variable and naming audits
Semantic tokens, aliases and modes
Properties, variants and state definitions
Figma-to-code mapping and usage guidance
AI-assisted build tests and correction review

See the design systems work ↗

Prototypes, frontend and tools

Make the behavior testable, then help carry the experience into the product.

Working code prototypes with representative inputs
Connected flows and realistic response states
Scenario tests, failure paths and design decisions
Frontend interactions and reusable components
Build review alongside product and engineering

Join at the point of need. Stay through delivery.

One senior person, complementary specialists or a problem for Less to own. We can join while the product is being defined, make the difficult behavior tangible, build the interface or improve the tools and systems behind it.

How we work together