By John Gordon, in partnership with Uncertainty Experts
Five articles. Three states. A thought leader landscape where Amodei, Huang, Georgieva, and the WEF disagree about pace but agree on one thing: skills are the differentiator.
Article 3 named Fear. Article 4 mapped Fog. Article 5 calculated the cost of Stasis. This article describes the structural answer. Not a tool recommendation. Not a pilot. A framework that resolves all three states and builds measurable capability.
The Five-Layer Model
Layer 1: Structured data (Markdown, Obsidian). Infrastructure. Article 2 described the Markdown convergence: every major agentic framework uses human-readable, vendor-neutral text. Layer 1 gives your team the same foundation the AI platforms themselves are building on.
Layer 2: AI Skills Catalogue (25 packs, 100 skills, SFIA 9). Capability. Major platforms converged on structured skills in 2025. Layer 2 organises them into AI Skill Packs mapped to SFIA 9's 147 skills across seven levels.
Layer 3: Governed version control (Git). Safety. Article 3 described how Fear presents as governance concerns. Layer 3 addresses them directly. Every piece of work is versioned, traceable, and auditable. Governance is built into the workflow, not layered on as a blocking function.
Layer 4: Practitioner pathway (user to orchestrator). Progression. Article 5 described the cost of Stasis. Layer 4 closes the gap with a defined path: user, practitioner, agent orchestrator. Every step is measurable and produces an artefact.
Layer 5: Leadership Labs. Strategy. For senior leaders who need to connect AI capability to business decisions without becoming practitioners themselves.
The Finer Journey
Assessment (free, 10 minutes). Surfaces which states are present and identifies which AI Skill Packs to prioritise. Diagnostic, not prescriptive.
Hands-On Sessions (£1,000 per delegate, two half-days, 15 to 20 delegates). Delegates build real AI skills with their own work materials. Not a demonstration. Not a presentation. They leave with artefacts they produced themselves.
Capability Programme (£150,000 per cohort, 10 to 12 delegates, three to six months). The full programme. Delegates work through AI Skill Packs mapped to their roles and build skills progressively. The programme does not teach tools. It builds capability. Tools change quarterly. Capability persists.
Leadership Labs (£10,000 per session). Connects AI capability to specific business questions: budgets, team structure, governance, return.
Retainer (negotiated). Ongoing access to new AI Skill Packs, updated frameworks, and strategic guidance as the landscape continues to shift.
Each step is independent. Each answers the question that prevents the next step from feeling premature. The Assessment answers "where do we stand?" Hands-On Sessions answer "can our people actually do this?" The Capability Programme answers "can we build this into how we work?" Leadership Labs answer "what does this mean for our strategy?"
The ROI case
We see between one and two days saved per person per week in our programmes. That is 8 to 16 hours, based on internal results. The published baseline without training is two to four hours.
The difference between two hours saved and 12 hours saved is not marginal. At 12 hours, an L&D professional spends more than a full day each week on strategic work previously consumed by production tasks. Benioff cited 30 to 50% productivity gains at Salesforce. PwC's wage premium data shows a 25% pay gap between AI-skilled and non-AI-skilled professionals. Both align with what we see in practice.
For 10 delegates over 26 weeks at a conservative 12 hours per person per week: 3,120 hours of recovered capacity. At £150,000, the cost per hour is approximately £48. Each delegate influences roughly five colleagues. Over six months, 10 delegates create measurable change across 50+ people.
Why this works when pilots do not
95% of AI pilots fail, according to MIT's research on the GenAI divide. The failure is structural, not technical. Pilots are tool-first: someone finds an impressive tool, runs a proof of concept, gets excited results for a narrow use case, discovers the results do not generalise.
The Finer Vision programme is framework-first. Every skill maps to SFIA 9. Every task maps to an AI domain. Every delegate is measured against the practitioner pathway. When a new tool appears, the framework absorbs it. SFIA 9 is vendor-neutral by design. Skills built now survive platform shifts.
Article 2 described how MCP, SKILL.md, and AGENTS.md are converging on common standards. The Finer Vision framework sits on top of that convergence. When a new standard appears, it maps to existing AI Skill Packs. When a new tool appears, the skills transfer. The framework is the constant. The tools are the variable.
What the end state looks like
The end state is an L&D function operating as agent orchestrators across four domains: human-only work (judgement, relationships), collaborative (AI assists, human decides), directed (human sets parameters, AI executes), and autonomous (AI operates within defined boundaries).
An L&D function operating across all four domains has not been replaced by AI. It has multiplied its capacity by a factor the organisation has never seen. The judgement, the organisational knowledge, the understanding of people: those are the scarce resources. AI handles the production. The L&D professional handles the thinking.
Research by LaFreniere and Newman found that 91.4% of worry predictions proved untrue. The catastrophic scenarios that keep teams in Stasis almost never happen. What does happen, consistently, is that teams who build skills discover they are more valuable than they were before. Not less.
Article 1 opened with a finding from UCL: uncertainty produces more stress than certain pain. There is a third option. Build the skills. Follow the framework. Measure the results.
The Finer Vision AI Maturity Assessment is free and takes 10 minutes. It is the first step. Understand where your team stands, which AI Skill Packs to prioritise, and whether you are ready for Hands-On Sessions, the full Capability Programme, or something in between.
Take the free AI Maturity Assessment at finervision.com/assessment