The Spark AI Maturity Model
The Spark AI Maturity Model™
What each stage looks like in practice – and the AI tools you'll see at each one
Stage 1
Experimentation
Stage 2
Adoption
Stage 3
Optimisation
Stage 4
Innovation
People
Individuals experiment in isolation
Shared training and language
Role-specific application
AI-native team culture
Process
No standards or governance
Acceptable use policies in place
AI embedded in workflows
Continuous process innovation
Data
No structured data use
Basic data literacy across roles
Data informs decisions
Predictive, real-time data use
Tools
Scattered, personal tool use
Approved tool stack defined
Tools integrated into delivery
Custom builds and automation
Strategy
No AI strategy exists
AI on leadership agenda
AI drives commercial decisions
AI is a competitive differentiator
AI Tools
You'll See
You'll See
- ChatGPT / Claude / Gemini / Copilot in blank chat windows
- Midjourney / Nano Banana experiments
- Fireflies / Otter for transcription
- Projects with brand voice & docs
- Custom GPTs / Gems for tasks
- Shared prompts on Slack
- AI for research & briefs
- Team-wide Skills & Projects
- Scheduled tasks (Cowork / Make / Zapier)
- Figma Weave for creative
- CRM connected to AI
- Multi-agent systems (Claude Code / Codex)
- Automated content & production pipelines
- Custom-built client tools
- AI informing strategy
How Spark
Helps
Helps
AI FundamentalsHalf-day workshop
AI Accelerator3-month programme
Strategic Coaching & Quarterly HorizonsOngoing advisory, custom tools, agent builds
Stage 1
Experimentation
People
Individuals experiment in isolation
Process
No standards or governance
Data
No structured data use
Tools
Scattered, personal tool use
Strategy
No AI strategy exists
AI Tools You'll See
- ChatGPT / Claude / Gemini / Copilot in blank chat windows
- Midjourney / Nano Banana experiments
- Fireflies / Otter for transcription
How Spark Helps
AI Fundamentals
Half-day workshop
Stage 2
Adoption
People
Shared training and language
Process
Acceptable use policies in place
Data
Basic data literacy across roles
Tools
Approved tool stack defined
Strategy
AI on leadership agenda
AI Tools You'll See
- Projects with brand voice & docs
- Custom GPTs / Gems for tasks
- Shared prompts on Slack
- AI for research & briefs
How Spark Helps
AI Accelerator
3-month programme
Stage 3
Optimisation
People
Role-specific application
Process
AI embedded in workflows
Data
Data informs decisions
Tools
Tools integrated into delivery
Strategy
AI drives commercial decisions
AI Tools You'll See
- Team-wide Skills & Projects
- Scheduled tasks (Cowork / Make / Zapier)
- Figma Weave for creative
- CRM connected to AI
How Spark Helps
AI Accelerator + Strategic Coaching
3-month programme & ongoing advisory
Stage 4
Innovation
People
AI-native team culture
Process
Continuous process innovation
Data
Predictive, real-time data use
Tools
Custom builds and automation
Strategy
AI is a competitive differentiator
AI Tools You'll See
- Multi-agent systems (Claude Code / Codex)
- Automated content & production pipelines
- Custom-built client tools
- AI informing strategy
How Spark Helps
Strategic Coaching & Quarterly Horizons
Ongoing advisory, custom tools, agent builds
Need some jargon busting?
| Term | What it means |
|---|---|
| Projects | A saved workspace inside your AI tool where you store instructions, upload reference documents, and keep context between conversations. Think of it as giving the AI a permanent briefing pack instead of starting from scratch every time. |
| Custom GPTs / Gems | Pre-configured AI assistants built for a specific task. You set the instructions and reference material once, then anyone on the team can use it. Custom GPTs are ChatGPT's version; Gems are Gemini's. |
| Skills | A newer feature (available in Claude, rolling out in ChatGPT) that packages a complete workflow into something you call by name. "Run the show notes skill on this transcript." One step beyond Custom GPTs – more structured, more reusable. |
| Scheduled tasks | An AI task that runs automatically on a set cadence – daily, weekly – without you triggering it. Like setting up an email rule, but for AI workflows. |
| Node-based workflows | A visual way of building AI processes where you connect steps together like a flowchart rather than typing prompts. Each "node" does one thing and passes the result to the next. Figma Weave is one example. |
| MCP connectors | The plumbing that lets your AI tool talk to other software like your CRM or email. You don't need to understand how it works – just know it's what makes "Claude, check my Monday.com board" possible. |
| Agents | AI that can reason about what needs to happen, choose the right tools, take action, and check its own work. Unlike a chatbot responding to one prompt at a time, an agent works through multi-step tasks more independently. |