The Team

One person. Six agents. Everything you see here.

This site was built by Lincoln Mitchell directing a team of six AI agents. Each has a defined role, clear boundaries, and real output you can inspect in this repo. This is what agentic design looks like in practice.

The Roster

Meet the team

Each agent has a dedicated context file in .claude/agents/defining its role, expertise, and boundaries. They don't freelance. They stay in their lane — and that's what makes them good.

CS

Lead Content Strategist

Writes everything in Lincoln's voice

ModelSonnet
BoundaryNever writes code
ProducesEvery word on every page
FD

Lead FED Developer

Builds what the team designs

ModelSonnet
BoundaryNever makes design decisions
ProducesEvery component on this site
UR

Lead UX Researcher

Researches the domain

ModelSonnet
BoundaryNever writes code or designs interfaces
ProducesDomain analysis that informs positioning
UI

Lead UI Designer

Owns the visual language

ModelSonnet
BoundaryNever writes production code
ProducesThe tonal surface system and typography
UX

Lead UX Designer

Owns how it works

ModelSonnet
BoundaryNever specifies colours or writes code
ProducesNavigation patterns and page structure
SS

Stakeholder Simulator

Pressure-tests ideas against real personas

ModelOpus
BoundaryNever recommends — only simulates
ProducesCTO, recruiter, and audience reactions
The agents execute. Lincoln decides. The human role is judgement, taste, and strategic alignment — exactly the things AI cannot self-generate.
The orchestration model
The Human in the Loop

The agents execute. Lincoln decides.

Lincoln is not a prompt engineer. He is a creative director running a team. He sets the vision, assigns tasks to the right agent, reviews output, corrects course, and approves the final result.

1Lincoln

Vision

Sets the goal and constraints

2Agent

Draft

Produces initial output

3Lincoln

Review

Corrects tone, strategy, accuracy

4Agent

Revise

Incorporates feedback

5Lincoln

Ship

Approves and commits

Governance

Why the Content Strategist doesn't write code

The boundaries are not limitations. They are the governance that makes the system work. When an agent stays in its lane, its output improves because the context window stays focused. Cross boundaries and quality drops.

Capability
Content
FED
Research
UI
UX
Stakeholder
Writes code
Writes content
Design decisions
Research
Light
Simulates personas
Reviews other agents
The Economics

What this team would cost at a consultancy

A traditional team to build a site of this quality with this depth of thought leadership content.

Traditional Team
UX Researcher£450–600/day
Content Strategist£400–550/day
UI Designer£450–600/day
UX Designer£450–600/day
Frontend Developer£500–700/day
Creative Director£700–1,000/day
5–6 people£2,950–4,050/day
The Agentic Team
1 Design Systems ArchitectLincoln
6 specialised AI agentsClaude
1 CLAUDE.md context fileThe system
One unified vision. One consistent voice.Zero alignment meetings.

This is not about replacing people. It is about what becomes possible when one person has the right methodology and the right tools. The agents handle volume and consistency. The human handles vision, judgement, and taste.

Want to build your own agent team?

The agent definitions are open source in this repo. The methodology is documented. The only thing you need is someone who knows how to orchestrate. That's what I do.

/