← Elements AI resources

Look Inside My Agentic AI Setup

The actual structures and prompts behind how I run Elements AI as an AI-native agency — the specialist agents, the skills they run, the memory they read from, and the config that ties it together. All real and in daily use; presented for you to explore and downloadable in a way your AI can consume.
Claire Byrne
From Claire — a personal note

The way I work with AI now is much more leadership-focused than ever before — communicating goals, intents and pain points — then we collaborate together on the solution and how it gets rolled out.

So what you’re looking at is my real setup, with client details stripped — what an agency owner who’s gone fully AI-native actually runs day to day. It won’t all fit you. Copy it wholesale and you’ll end up carrying pieces you don’t need — which rather misses the point. It’s here to give you and your AI something concrete to react to, so you can shape a way of working that’s yours.

I rebuilt how I work day-to-day about five months ago to suit the intelligence available then, and this is only a handful of the concepts — there are more layers underneath. It also keeps evolving: the latest models (like Fable) can infer a lot without being explicitly briefed, so some of what I do today will be redundant, or even counterproductive, within months — while other parts hold. So reassess it at every step up in intelligence, and take the thinking, not the literal setup — use it to build your own AI collaboration style, and your own AI team.

Any questions or feedback, I’d love to hear from you — reach me on LinkedIn or email.

Claire
Two ways to use this
Everything here is based on my real setup, anonymised and adapted so your AI can help you implement a tailored version in your workspace. Take the guided tour and copy what you want piece by piece, or grab the whole thing as one pack and let your own Claude Code install it with you.
A
Walk it
Step through what I've built, in order — the AI stack underneath, the org memory that keeps it from guessing, the team of specialist agents, and the CLAUDE.md that sets the rules. Copy or download any piece as you go.
  • AI Stack — the four-layer operating model
  • Memory — the org memory it reads from
  • Roles — 9 specialist agents
  • CLAUDE.md — the config that shapes behaviour
B
Run it
Don't want to do it by hand? Download one pack and hand it to your own Claude Code. It checks what you already run, works out which of these actually fit your business, and helps you install just those — no mass dump.
  • Audits your current setup first
  • Scores each role & skill for relevance to you
  • Recommends a shortlist — you approve each one
  • Installs and wires up only what you said yes to
For Claude Code today (a Cowork version is next). Hand it the file and say "set this up for me."
Keep going
CB

These are the specialist AI agents I've set up — I think of them as coworkers. Each is a persona with its own job, voice and judgment, and the skills are the actual tasks they run.

It's a "skilled staff base" — a team that shares the same context, not a single all-purpose assistant.

9 specialist agents
Each one is a specialist AI agent — a distinct senior persona that owns its area and works in its own voice. Tap any agent to view its skills and a sample prompt.
↑ Tap any agent to open it
Skills any role can pick up
Cross-cutting skills available to every role, grouped by purpose.
Yours to keep — the full skill, handed over Reference only — shown for context
Keep going
CB

This is the operating model behind everything else: the four tools that turn a chat window into a coworker. Pick one tool per layer and you've got a working setup.

  • Workspace — where the AI actually does the work
  • Memory — the files and notes it can read and own
  • Ears — captures my meetings and calls (this is "bringing your agent into the room")
  • Voice — so I can talk instead of type

The starred tool is what I use today; everything under it is a swap-in alternative. The shape doesn't change when a tool does.

Four layers, one tool each
Each layer below shows the tool I use, starred, with the alternatives I'd swap in underneath it.
Workroom & LLM
Where the AI does the work
Candidates
Memory
Files it can read & own
★ Notion
Candidates
Ears
Captures meetings & calls
Candidates
Voice
Talk instead of type
Candidates
  • Whisper
  • LLM tooling
  • Computer dictation
Keep going
CB

Your AI needs context — and yours is scattered.

Good work from an AI depends on context: what you've done, what you've made, and what you've decided on as a team. But that usually lives in lots of different places — files on disk, the AI tool's own memory, wikis and drives — and the AI can't reliably reach into any of them. So it guesses. And a guessing agent hallucinates, burns tokens hunting, and hands you work that isn't business-ready.

Pull it into one organised place, and you get three things back:

Step 1 · The home
Choose a home for your files
Fixing this is an information-architecture problem, not a tooling one. Three things make a store your AI can actually use:
✓
A clear home
Decide where things are stored, and keep it to one place — so your AI never has to guess where to look.
✓
A categorization system
Organise it logically, so related things sit together and there's an obvious spot for everything that comes in.
✓
Way of working documented
Work with your AI to examine your file home(s), so you're both able to fetch exactly what a task needs, quickly.
Demo · Claire's Setup
Workspace Home
I've chosen Notion + Files as where my brain lives, as it gives me cloud accessibility as well as an editing interface I can jump into and share from.
Client Client Client Project (+ rollup) Related Projects Clients the nucleus Meeting Summaries the conversations Projects the spine Resources the knowledge Tasks next actions
Demo file structure · tap any database for its columns and joins
The one rule that makes it work: every record points back to a Client. That single discipline turns five tables into one memory the agent can walk.
Step 2 · The instructions
Provision it yourself
A single instruction you paste into an AI connected to your workspace. New to this? It builds the five-database structure for you to fill. Already have a system (a workspace or CRM)? It audits what you've got and proposes the lightest changes to get the same benefits — it won't rebuild what you already have.
Keep going
CB

This is the config that shapes how your AI behaves — your CLAUDE.md, the standing instructions it loads at the start of every session. It's written for the AI rather than for people, so it rarely gets read — and that's why most are thin. The aim of this tab is simple: get yours to look like this. Copy the prompt below, give it to your Claude Code or Cowork, and let it find the gaps in your current config, ask you what it needs, and rework it with you.

The coverage map
A mature CLAUDE.md tends to cover these. Use it as the target shape — the prompt below checks your file against it.
Hard rules
The non-negotiables the AI must never break.
About the business
What you do, who you serve, how you position.
About you
Your role, how you work, communication style, what to avoid.
Session startup
What the AI should do before any work begins.
Source of truth
Where operational data lives vs what stays in config.
Tools & connectors
Which integrations are connected, and the rules for each.
Roles
The specialist personas available and what each owns.
Skills
The reusable workflows and when to use them.
Model notes
Which model or tier to use for which work (optional).
Upgrade your config in one paste
Copy this into Claude Code or Cowork. It reads your current CLAUDE.md, scores it against the map above, asks you what it needs, then proposes a reworked version — so yours ends up much closer to this.
One more
CB
Note from Claire

If you’ve worked through these tabs with your AI, you’ve built the bones of an operating model: a coworker with a memory, somewhere to keep its context, and standing instructions — not a chatbot you re-brief every morning. You won’t have all of it, and you shouldn’t: take the pieces that fit how you actually work and leave the rest.

Here’s the shape of what “done” looks like — the four things you’ll have built, and what they give you back. Tick what’s true for you now.

Claire
Stages 1 to 4 checklist
✓
STAGE 1
I have an AI I work inside, not just chat with.
✓
STAGE 2
I have my business context in one place my AI can reach.
✓
STAGE 3
I have standing instructions — my CLAUDE.md — so my AI starts every session knowing how I work.
✓
STAGE 4
I have my meetings captured and filed automatically, so the thinking behind decisions isn’t lost.
Capability unlocks · I can now…
✓
Get business-ready work back the first time.
✓
Correct my AI once, and it holds.
✓
Walk into any piece of work and get oriented in seconds, not an hour of digging.
✓
Keep the reasoning behind every meeting — the why, not just the notes.
✓
Pick up work from my phone with my full context loaded.
✓
Hand off a task and trust the output reflects how we actually operate.
That’s the tour
×