AI Value Realisation Pathway
From reactive assistance to workflow transformation — a 4-stage framework for AI value that compounds. Where most organisations plateau, and what it takes to break through.
Note
Updated Framework Available: This article presents our original 4-stage framework (March 2026). For the expanded 5-stage framework incorporating agentic AI and governance, see Five Stages to Compounding Returns.
88% of organisations use AI. Only 6% capture meaningful enterprise-level value.
The technology isn't the problem. The way it's deployed is.
Most AI investments follow the same pattern: deploy a tool, wait for value, wonder why it didn't arrive. The issue isn't the model, the vendor, or the budget. It's that AI value doesn't come from AI — it comes from re-engineering the work that AI is meant to improve. That is a fundamentally different problem, and most organisations aren't solving it.
This framework maps four stages of AI value realisation — from ad-hoc tool usage to architected workflow transformation. Understanding which stage your organisation is at is the foundation of any credible AI strategy.
Reactive Assistance
When Every Question Starts Fresh
Structured Experimentation
Building Blocks Emerge
Connected Intelligence
AI Meets Your Data
Workflow Transformation
The Architectural Shift
Most organisations plateau at Stages 1–2. The leap to Stage 4 requires separating what's predictable (codify it) from what's creative (AI enhances it).
The stage progression is not about AI sophistication. It is about operational maturity. A Stage 1 organisation and a Stage 4 organisation might use the same AI models, the same APIs, the same vendors. What differs is how the work is structured around those models — whether knowledge is captured and scaled, whether workflows have been redesigned or just assisted, and whether value compounds or resets every time someone opens a new chat window.
Understanding Each Stage in Practice
The overview above gives you orientation. What follows is the operational reality of each stage — what it looks and feels like from inside the organisation, what it costs to stay there, and what it actually takes to move forward.
Select any stage below to explore it in depth.
Your teams use AI when they remember to. Someone pastes a document into ChatGPT. Another rewrites the same prompt they wrote last week. Knowledge stays in individual heads, and every interaction starts from zero.
You're at this stage if...
- "Just ask ChatGPT" is common advice in meetings
- No one knows who is paying for AI subscriptions
- The same prompts get reinvented weekly
- Success stories are anecdotal, not measured
The Hidden Cost
At a 50-person company, reactive AI usage wastes ~120 hours/month in prompt recreation alone. Teams are spending time using AI without any system to capture and scale what works.
The Breakthrough
Document what works. Create a shared prompt library. Identify your top 5 use cases. This alone typically delivers 5–15% individual productivity improvement.
"We use AI" means nothing without a system to capture and scale what works.
Stage 1 Reality Check
Teams begin saving prompts and building reusable projects. Champions emerge who share what works. But token costs spiral, knowledge stays siloed, and scaling feels impossible.
You're at this stage if...
- You have Claude Projects or Custom GPTs in use
- Some teams have documented their AI workflows
- Token costs are tracked but not optimised
- Success depends on specific individuals
The Hidden Cost
Organisations at this stage typically spend 3x more on tokens than necessary due to inefficient prompting. When your AI champion leaves, the knowledge walks out the door.
The Breakthrough
Standardise on 2–3 platforms. Create team templates. Measure time-to-value, not just usage. This is where you start building repeatable capability.
"AI champions" are a liability if their knowledge cannot be institutionalised.
Stage 2 Reality Check
AI tools now connect to your email, documents, and calendars. Context is richer. But you are still asking AI to do the same manual workflows — just with better information. The ceiling is visible.
You're at this stage if...
- Copilot or Gemini is deployed across the org
- AI can access internal documents and emails
- People use AI as "super search"
- Workflows themselves have not changed
The Hidden Cost
Connected AI without workflow change delivers only 60% of potential value. The rest requires architectural thinking — separating what's predictable from what's creative.
The Breakthrough
Map your workflows. Identify what is predictable vs creative. Begin separating the two. This is where you stop adding AI to existing processes and start redesigning the processes themselves.
"Connected AI" is still manual AI. Access is not architecture.
Stage 3 Reality Check
The breakthrough stage. Predictable work is codified into applications — business rules in code, not in prompts. AI APIs handle creative judgment: pattern recognition, document understanding, exception identification. Humans oversee decisions, governance, and true edge cases. Value compounds because the architecture is designed for it.
You're at this stage if...
- Custom platforms handle routine operations
- AI is called via APIs for specific tasks
- Deterministic logic is in code, not prompts
- ROI is measured and growing
The Hidden Cost
Without reaching Stage 4, organisations leave the majority of their potential AI value unrealised. The gap isn't technology — it's the willingness to re-engineer how work gets done.
The Breakthrough
This is not a technology project — it is a business architecture project. Start with one workflow. Map it. Separate predictable from creative. Codify the predictable. Augment the creative with AI.
"The goal is not AI everywhere — it is AI where creativity matters."
Stage 4 Reality Check
The movement from Stage 1 to Stage 3 is largely additive — more tools, better prompts, richer context. The movement from Stage 3 to Stage 4 is categorically different. It requires a deliberate decision to stop optimising existing workflows and start redesigning them. Most organisations are not willing to do this, which is why the majority plateau indefinitely at Stage 2–3.
The ones that break through share a common insight: predictable work belongs in code, not in prompts. AI should handle what genuinely requires judgment — not serve as an expensive substitute for conditional logic that a developer could codify in an afternoon.
Where Does Your Organisation Stand?
Honest self-assessment is the foundation of effective AI strategy. Most organisations overestimate their maturity — the presence of AI tools is mistaken for AI capability. The diagnostic below maps your current state across the five dimensions that most reliably predict where value is being realised and where it is being left on the table.
Where Is Your Organisation?
Answer 5 questions to identify your current stage.
How do your teams primarily use AI today?
Your result is a starting point, not a ceiling. Most organisations that have reached Stage 4 did not arrive there in a single initiative. They built incrementally — one workflow at a time — until the architecture became self-evident and the compounding returns became visible enough to justify the next step.
What the Research Confirms
The patterns we observe working directly inside operational workflows are consistent with what the leading research houses report across thousands of enterprise AI deployments. This is not isolated experience — it is a structural pattern.
What the Research Confirms
Consistent findings from McKinsey, BCG, and Gartner — not opinions.
of AI initiatives in infrastructure and operations fully meet ROI expectations — meaning 72% stall or fail
Gartner, April 2026more likely to achieve significant EBIT impact when workflows are fundamentally redesigned around AI
McKinsey, 2025of AI transformation value comes from people and process change — not from algorithms or the model itself
BCG, 2024Why Most Organisations Plateau
The gap between Stage 2–3 and Stage 4 is not technological — it is architectural. Organisations try to layer AI onto existing workflows instead of redesigning those workflows for an AI-native operating model. The solution is not more AI tools. It is separating what is deterministic (build it in code) from what requires judgment (AI excels here) — then engineering each layer for what it does best. This is not a technology project. It is a business re-engineering project.
What It Takes to Move Forward
The jump from experimentation to transformation requires three things operating together. Most organisations attempt one or two. The ones that achieve Stage 4 build all three in sequence.
Workflow re-engineering. Understanding the work deeply enough to separate what is predictable from what requires creative judgment. This cannot be done from a distance. It requires embedding inside the operational workflow, watching how decisions are actually made, and identifying where deterministic logic is hiding behind manual steps.
Application architecture. Codifying that deterministic logic into reliable, auditable software. Business rules in code, not in prompts. This is what makes automation fast, predictable, and capable of operating without constant human oversight. It is also what prevents the token cost spiral that kills Stage 2–3 AI programmes.
Operational engineering. Building governance, exception handling, monitoring, and continuous improvement into the system from day one. Production AI is not a project — it is a system that requires daily operational ownership. The handoff model that works for traditional software projects fails for AI-augmented workflows, because the edge cases accumulate and the model behaviour shifts.
This is what we do. We embed inside operational workflows, re-engineer them from first principles, build the technology layer around what we find, and then continue to operate it. We live this ourselves — our 120-person team across five regions runs entirely on purpose-built internal platforms that follow exactly this architecture.
The staying is the differentiator. Not because it is clever positioning, but because production intelligent automation genuinely requires ongoing operational engineering to sustain and compound its value.
Start with one workflow. Book a conversation to map it, separate the predictable from the creative, and see exactly where the value sits.
Sources
-
McKinsey & Company. The State of AI in 2025: Agents, Innovation, and Transformation. McKinsey QuantumBlack, November 2025. Survey of 1,993 respondents across 105 countries. mckinsey.com
-
Gartner. AI Projects in I&O Stall Ahead of Meaningful ROI Returns. Press release, 7 April 2026. Survey of 782 infrastructure and operations leaders. gartner.com
-
Boston Consulting Group. Where's the Value in AI? BCG, 2024. Global study of 1,000 organisations. bcg.com
-
Boston Consulting Group. The Widening AI Value Gap. BCG Build for the Future, October 2025. Global survey of 1,250 senior executives and AI decision-makers. bcg.com
Start with one workflow.
Map it. Separate predictable from creative. See exactly where AI adds value — and where it doesn't.