How It Works
This page explains how SkillsLayer AI works -- from subscribing and activating skills to using the AI Brain's routing, chaining, and memory systems. No technical knowledge required.
What is a skill?
A skill is a specialist instruction module that lives inside your SkillsLayer AI account. When you activate a skill, the AI Brain automatically loads it before every relevant response -- giving your AI specialist knowledge, defined workflows, and consistent output formats for that domain.
Think of it as briefing a contractor before every job, except the briefing happens automatically and it never forgets.
Skills contain domain expertise, prompt frameworks, output templates, and routing logic. They run entirely inside your SkillsLayer AI account and work with any AI provider -- DeepSeek, Groq, OpenAI, Anthropic, and more.
Skills do not send your data anywhere. Everything runs inside your SkillsLayer AI account under our standard privacy policy.
Requirements
To get started you need:
- A SkillsLayer AI subscription -- plans from £19.99/month give you access to the full skill library and AI Brain.
- Your own API key from any supported AI provider (DeepSeek, Groq, OpenAI, Anthropic, Mistral, and more).
Installing a pack
Installation takes under two minutes and requires no technical knowledge. All steps happen through the web interface.
Subscribe and log in
Choose a plan from £19.99/month at app.skillslayerai.com. Your account includes the AI Brain, persistent memory, and access to the full skill library.
Activate your skills
Browse the skill library from your dashboard and activate up to your plan limit. Skills are organised by domain -- SEO, HR, e-commerce, operations, real estate, and more. Swap them anytime.
Connect your AI provider
Go to AI Settings and paste in your API key from any supported provider -- DeepSeek, Groq, OpenAI, Anthropic, Mistral, xAI, and more. Your key stays secure in your account.
Create a client workspace
Click New Client on your dashboard and give it a name. Each client gets their own Brain, memory layer, and conversation history -- completely separate from your other clients.
Onboard the client
Open a chat and type brain, new client. The Brain walks you through capturing their business details, brand voice, goals, and constraints into persistent memory.
Start working
Describe what you need in plain language. The Brain routes to the right skill automatically, chains multiple skills for complex tasks, and saves what it learns at the end of every session.
Working with multiple clients
Each client workspace has its own AI Brain, memory layer, and conversation history. The Brain reads the client's context at the start of every session automatically. You can have as many client workspaces as your plan allows.
One workspace per client (recommended)
Create a separate workspace for each client from your dashboard. Each workspace has its own memory, conversation history, and skill context. Everything stays completely separate -- no risk of one client's details affecting another's work.
This is the cleanest setup. Use it if your clients have very different brands, industries, or if you want isolated memory per client.
Click New Client on the dashboard
Name it after the client -- e.g. "Hartley Digital". The Brain and memory layer are set up automatically.
Run the onboarding command
Open a chat in the workspace and type brain, new client. The Brain captures their business details, brand voice, goals, and constraints into persistent memory.
Run brain, new client [name]
Start a chat in the Project and type brain, new client [client name]. The Brain walks you through capturing any missing context and confirms what it now knows about them.
Multiple clients in one workspace
You can also name the client in your message and the Brain will focus on their context within the same workspace. Useful for quick tasks across similar clients.
Name the client in your request -- "Brain, write a case study for Acme Co" -- and the Brain focuses on that client. Save to memory at the end of the session to capture what was learned.
Your first session
Once you are in a client workspace, just describe what you need in plain language. You do not need to name a skill -- the Brain routes to the right one automatically.
You: Run an SEO audit on this page. Here is the URL and the content...
You: Write a case study for my client, Acme Co. They are a logistics company. The result they want to highlight is a 30% reduction in delivery times.
If you have the AI Brain, you can also use the brain command prefix to invoke the orchestrator directly:
Brain modes
The AI Brain has four operating modes. It selects the right one automatically based on your request, but you can also state which mode you want explicitly.
Router mode
Matches your request to a single skill and executes it. Best for specific, well-defined tasks: "write a press release", "run a keyword gap analysis".
Chainer mode
Executes multiple skills in sequence to complete a complex task. Best for multi-step workflows: "plan and write a content campaign for this product launch".
Strategist mode
Open-ended planning and roadmapping. Draws on multiple ecosystems to produce strategic output without following a fixed skill sequence.
Learner mode
Focuses on capturing new client information, updating memory, and refining skill patterns. Use it to onboard a new client or after a major project.
Memory layer
Every standard AI session starts from scratch. The Brain does not. It carries forward everything it knows about your business, your clients, and what has worked — so each session picks up where the last one left off.
Before your first session, run the onboarding command in your client workspace. The Brain captures everything it needs into persistent memory and loads it automatically at the start of every session from that point on.
As you work, the Brain quietly builds on that foundation — noting which approaches land well, spotting gaps in its own skill coverage, and learning how you like outputs shaped. None of that disappears when you close the chat, as long as you save at the end.
Closing a session
When you are done, click Save to Memory. The Brain reviews what was learned in the session and automatically updates the relevant memory files. You can review the changes before confirming.
Only memory files that actually changed are updated. A short single-skill session might change just one file, or none at all. Click Save to Memory and the Brain handles it automatically.
A typical working week
Here is what using the Brain actually looks like in practice across a normal week of client work. No scripts, no special commands to remember — just describe what you need and the Brain handles the rest.
Start a session — Brain loads your context
Open your workspace and start a chat. The Brain reads your client context and memory files automatically. It already knows who your clients are, what you are working on, and what has worked before. No re-explaining needed.
Ask for what you need — Brain routes and executes
"Brain, write this month's newsletter for Hartley Digital. Focus on the new product launch. Tone should match their brand voice." The Brain picks the right skill, pulls the client context, and delivers a shaped draft — not a generic template.
End the session — ask for a memory update
Click Save to Memory at the end of the session. The Brain updates what it learned automatically. Next session it already knows the newsletter was written, what approach you used, and what the client's preferences are.
Chain a multi-step workflow
"Brain, run an SEO audit on this page and turn the findings into a prioritised action plan." The Brain chains two skills in sequence — audit first, action plan from the results — without you managing the handoff. One request, complete output.
End-of-week review
"Brain, status." The Brain surfaces your five strongest skills for this client, your top gaps, and a recommended focus area for the next 30 days — based on everything it has learned across the week's sessions.
New client? Onboard in minutes
Create a new client workspace on the dashboard and run brain, new client. The Brain captures their details automatically. Done -- it knows who they are from the first message.
After a month of regular sessions the Brain knows your top clients inside out, has a running log of every gap it has hit, and routes requests with real context behind every decision. It is not the same tool it was on day one — and it keeps improving the more you use it.
Single skill examples
Each skill is self-contained and can be triggered with plain language. Here are examples across the 17 skill ecosystems:
SEO & Search
"Give me a keyword research report for a UK-based plumbing company targeting emergency call-out work in Cardiff."
"Run a full SEO audit on this page. Here is the HTML and the target keyword: 'best accountants Birmingham'."
Content Marketing
"Write a case study for Greenleaf Roofing. The result: they reduced their quote-to-close time from 14 days to 3 days after using our CRM template."
"Write this month's newsletter for my accountancy client. Focus: the new R&D tax credit changes. Tone: professional but approachable."
Client Retention
"Generate an invoice for Hartley Digital. Scope: website redesign, 3 phases. Total: £4,800. Payment terms: 50% upfront, 50% on completion."
"Create an onboarding checklist for a new social media management client. They are a restaurant group with 4 venues."
Analytics & Reporting
"Write a monthly analytics report for my client using these GA4 figures. [paste figures]. Format it for a non-technical audience."
Chaining skills
Chaining means the Brain executes multiple skills in sequence, passing output from one into the next. You trigger it by describing a multi-step goal rather than a single output.
Full workflow examples
These show end-to-end tasks the Brain handles across multiple sessions:
New client onboarding
- Create a new client workspace and run the onboarding command.
- Run: "brain, create an onboarding checklist for [client name]. They are a [business type]."
- Run: "brain, generate a project brief for the [project name] scope we discussed."
- Run: "brain, write a welcome email for [client name] confirming the start date and next steps."
- Run /save to capture the new client context in memory.
Monthly business report
- Paste the month's GA4 and social metrics into the chat.
- Run: "brain, produce this month's full business report for [client name] using these figures."
- The Brain chains: channel-reporting + analytics-report + kpi-dashboard summary.
- Run /save to record performance patterns for next month's comparison.
SEO sprint
- Run: "brain, run a full SEO audit for [domain]. Priority: local search and page speed."
- The Brain chains: page-seo-audit + page-speed-audit + local-seo + metadata-audit.
- Run: "brain, turn those audit findings into a prioritised action plan."
- Run /save so the audit findings are stored for follow-up next session.
Commands
Every command below works on its own or chained with individual skills using brain, chain [skill+skill]. All commands are typed directly into the chat inside your workspace.
Brain commands
Built-in workflow chains
These are pre-built multi-skill sequences you can trigger with a single command. Each one runs the skills in order, passing output through the chain automatically.
brain, launch [thing]. Covers every channel from positioning to launch day assets.Building your own chains
Any skill in the index can be chained with any other. Use brain, chain [skill+skill+skill] to run a custom sequence, or just describe the end goal and let the Brain build the chain itself.
You: brain, chain keyword-research+content-brief+blog-content-template for a post targeting "best CRM for freelancers UK"
You: brain, plan — I want to build authority in local SEO for my plumbing clients over the next 3 months
Individual packs
Individual packs are self-contained and do not require a memory layer to work. You can add more packs to the same Project at any time - each one adds a new domain of expertise.
Upgrading to the full Brain at any time adds the orchestrator and memory layer on top of your existing packs.
FAQ
Can I use the Brain with multiple clients?
Yes — two ways. The cleanest is one workspace per client, created from your dashboard. Each has its own memory and conversation history completely isolated. Alternatively, you can name the client in your message and the Brain focuses on their context within the same workspace. Use brain, new client to onboard each one.
Can I use skills on the free plan?
No - AI workspaces, where skills are stored, require a paid plan (Pro, Team, or Enterprise).
Do skills work with advanced AI models as well as Sonnet?
Yes. Skills are optimised for the latest AI model, but they work with any AI model available in your Project settings. Opus will produce higher quality output on complex chains at higher cost per session.
Will updates to my purchased packs always be free?
Yes. All updates to any pack you have purchased are free for life. Check the Updates page for the latest changelogs and download new versions from your purchase confirmation email or by contacting us.
Can I use the same pack across multiple AI workspaces?
Yes. Your licence allows you to install the pack in as many Projects as you need for your own use. You cannot share or resell the files.
What if a skill does not cover what I need?
The Brain logs gaps automatically during sessions - run /save and check the gap-log.md output. You can also contact us to request a new skill or ecosystem.