YARN Studiopack-driven assessments · Agile Insights · v0.1.0

AI Maturity (Agile Yarn) v2.0.0

ai-maturity · 6 sections

⚡ Generate from source material ⬇ Export .yarnpack.json Advanced: raw JSON

Metadata

Scales, categories & audiences

Scales: maturity5 (5 levels) · Categories: strategy value org people gov eng data · Audiences: exec data gov people biz

Edit these in the raw JSON editor below — they change rarely and carry structure (levels, signals, deep-dive wiring) that a form would mangle.

Core capabilities core

Everyone answers these.

Id	Question	Type	Category	Scale
vision	How clear is your organisation's vision and ambition for AI? Is there a documented AI strategy, and who owns it?	scored_text · scored	strategy	maturity5
roadmap	Is there a roadmap for AI with short- and long-term priorities? How are AI investments decided and sequenced?	scored_text · scored	strategy	maturity5
prioritise	How do AI use cases get identified, evaluated and prioritised? Walk me through how a new idea becomes a funded project.	scored_text · scored	value	maturity5
businesscase	How do you build the business case for an AI initiative, and how is value actually measured after it ships?	scored_text · scored	value	maturity5
operatingmodel	How is AI organised — central team, embedded, federated? Is there a defined operating model, or is it improvised per project?	scored_text · scored	org	maturity5
culture	How would you describe the organisation's appetite for AI? How is adoption and change managed?	scored_text · scored	people	maturity5
skills	How well do you understand the AI skills you have versus what you need? Is there a plan to close the gaps?	scored_text · scored	people	maturity5
training	What AI training, upskilling or literacy programs exist, and who actually has access to them?	scored_text · scored	people	maturity5
ethics	What policies or principles govern responsible and ethical use of AI? Are they actually enforced, or shelfware?	scored_text · scored	gov	maturity5
risk	How are AI-specific risks identified and managed? Is there a risk framework or register for AI?	scored_text · scored	gov	maturity5
compliance	How do you monitor and enforce compliance with AI policies and legal/regulatory obligations?	scored_text · scored	gov	maturity5
architecture	Describe the platforms and infrastructure you use to build and run AI. How standardised is it across teams?	scored_text · scored	eng	maturity5
deploy	Once an AI solution is built, how does it get deployed, monitored and supported in production?	scored_text · scored	eng	maturity5
dataacquire	How easily can teams get the data they need for AI? Walk me through sourcing and preparing it.	scored_text · scored	data	maturity5
dataquality	How confident are you in the quality, completeness and consistency of data used for AI?	scored_text · scored	data	maturity5
datagov	What governance exists over data used in AI — ownership, access controls, lineage?	scored_text · scored	data	maturity5
datasec	How is sensitive data protected when used in AI? Any privacy controls specific to AI?	scored_text · scored	data	maturity5

＋ Add question to “Core capabilities”

Strategy deep-dive strategy_dd · audience: exec

Extra depth for the areas you know best.

Id	Question	Type	Category	Scale
trendscan	How do you monitor and interpret emerging AI trends and assess their impact on your business?	scored_text · scored	strategy	maturity5
readiness	Do you formally assess organisational readiness for AI? How?	scored_text · scored	strategy	maturity5
innovation	How are new AI ideas and experiments fostered, captured and scaled?	scored_text · scored	value	maturity5
ecosystem	How do you use external partners, vendors and alliances in your AI strategy?	scored_text · scored	org	maturity5

＋ Add question to “Strategy deep-dive”

Data deep-dive data_dd · audience: data

Extra depth for the areas you know best.

Id	Question	Type	Category	Scale
dataprep	How mature are your data preparation and feature-engineering practices for AI?	scored_text · scored	data	maturity5
dataintegration	How well integrated are your data sources for AI use, internal and external?	scored_text · scored	data	maturity5
lineage	Can you trace the origin and transformation of data feeding AI models?	scored_text · scored	data	maturity5
metadata	What metadata, cataloguing or semantic-layer capabilities support AI?	scored_text · scored	data	maturity5
observability	How do you monitor the health and behaviour of data over time?	scored_text · scored	data	maturity5
analytics	How are analytics and visualisation used to support AI insight?	scored_text · scored	data	maturity5

＋ Add question to “Data deep-dive”

Engineering deep-dive eng_dd · audience: data

Extra depth for the areas you know best.

Id	Question	Type	Category	Scale
buildvsbuy	How do you decide build vs buy for AI capabilities? Is there a rule or is it case-by-case?	scored_text · scored	eng	maturity5
buildapps	Describe how AI applications and models actually get built and trained here.	scored_text · scored	eng	maturity5
testing	How are AI models and applications tested before and after release?	scored_text · scored	eng	maturity5
techdebt	How do you track and manage technical debt in AI systems?	scored_text · scored	eng	maturity5
scale	How do you scale AI systems that prove valuable?	scored_text · scored	eng	maturity5
mlops	What MLOps practices keep AI running in line with service levels?	scored_text · scored	eng	maturity5

＋ Add question to “Engineering deep-dive”

Governance deep-dive gov_dd · audience: gov

Extra depth for the areas you know best.

Id	Question	Type	Category	Scale
security	What frameworks address AI-specific security and safety threats?	scored_text · scored	gov	maturity5
legal	How do you track and respond to AI legal and regulatory requirements?	scored_text · scored	gov	maturity5
sustainable	Are sustainability and cost-of-AI considerations governed at all?	scored_text · scored	gov	maturity5
govteams	Who governs AI? Are there cross-functional teams and clear decision rights?	scored_text · scored	gov	maturity5
oversight	What systems provide continuous oversight and assurance of AI in production?	scored_text · scored	gov	maturity5

＋ Add question to “Governance deep-dive”

People deep-dive people_dd · audience: people

Extra depth for the areas you know best.

Id	Question	Type	Category	Scale
workforce	How do you forecast and plan for AI's impact on roles and staffing?	scored_text · scored	people	maturity5
recruit	How do you attract and recruit AI talent?	scored_text · scored	people	maturity5
trainingeff	How do you measure whether AI training actually changes anything?	scored_text · scored	people	maturity5
changemgmt	How are AI-driven change initiatives executed and monitored?	scored_text · scored	people	maturity5

＋ Add question to “People deep-dive”

Add section

Advanced — raw JSON

Full pack document, validated on save (schema yarn-pack/2). This is where scales, audiences, structured conditions and adaptive config live.

{
  "schema": "yarn-pack/2",
  "id": "ai-maturity",
  "version": "2.0.0",
  "name": "AI Maturity (Agile Yarn)",
  "engagement": "AISG — AI maturity pre-read / front-door",
  "intro": "A short, guided conversation — not a form. Answer in your own words; attach a document if it helps. Takes 15–25 minutes.",
  "tone_default": "professional",
  "prefill_fields": {
    "department": [
      "Executive",
      "Finance",
      "Operations",
      "Customer / Sales",
      "Technology / IT",
      "Data & Analytics",
      "People & Culture",
      "Risk & Compliance",
      "Marketing",
      "Product"
    ]
  },
  "scales": [
    {
      "id": "maturity5",
      "name": "Maturity (distilled model)",
      "levels": [
        {
          "value": 1,
          "label": "Does not exist",
          "gloss": "No capability. Absent, or purely ad hoc / accidental."
        },
        {
          "value": 2,
          "label": "Partially exists",
          "gloss": "Emerging and inconsistent. Pockets of activity, not joined up."
        },
        {
          "value": 3,
          "label": "Fully exists",
          "gloss": "Defined, documented and operating across the organisation."
        },
        {
          "value": 4,
          "label": "Fully exists & optimised",
          "gloss": "Measured, refined and improving against targets."
        },
        {
          "value": 5,
          "label": "Fully exists & adaptive",
          "gloss": "Continuously self-adjusting; a source of advantage."
        }
      ],
      "signals": {
        "1": [
          "no ",
          "not ",
          "none",
          "never",
          "don't",
          "do not",
          "nothing",
          "absent",
          "unaware",
          "haven't",
          "ad hoc",
          "ad-hoc",
          "nonexistent",
          "no idea",
          "not really"
        ],
        "2": [
          "some ",
          "starting",
          "beginning",
          "emerging",
          "pilot",
          "trial",
          "informal",
          "inconsistent",
          "pockets",
          "a bit",
          "occasionally",
          "early",
          "experiment",
          "trying",
          "patchy"
        ],
        "3": [
          "documented",
          "defined",
          "standard",
          "standardised",
          "established",
          "policy",
          "framework",
          "process",
          "consistent",
          "across the",
          "in place",
          "formal",
          "governed",
          "rolled out"
        ],
        "4": [
          "measured",
          "metrics",
          "optimis",
          "improving",
          "kpi",
          "monitored",
          "reviewed",
          "refined",
          "benchmarked",
          "targets",
          "tracked",
          "mature",
          "regularly review"
        ],
        "5": [
          "continuous",
          "adaptive",
          "self-",
          "automated end",
          "best in class",
          "best-in-class",
          "competitive advantage",
          "industry leading",
          "always",
          "real-time monitoring",
          "feedback loop"
        ]
      }
    }
  ],
  "categories": [
    {
      "id": "strategy",
      "name": "AI Strategy",
      "order": 1,
      "target_default": 3
    },
    {
      "id": "value",
      "name": "AI Value",
      "order": 2,
      "target_default": 3
    },
    {
      "id": "org",
      "name": "AI Organisation",
      "order": 3,
      "target_default": 3
    },
    {
      "id": "people",
      "name": "AI People & Culture",
      "order": 4,
      "target_default": 3
    },
    {
      "id": "gov",
      "name": "AI Governance",
      "order": 5,
      "target_default": 3
    },
    {
      "id": "eng",
      "name": "AI Engineering",
      "order": 6,
      "target_default": 3
    },
    {
      "id": "data",
      "name": "AI Data",
      "order": 7,
      "target_default": 3
    }
  ],
  "audiences": [
    {
      "id": "exec",
      "name": "Executive / Strategy",
      "desc": "Leadership, strategy, investment",
      "deep_dive_sections": [
        "strategy_dd"
      ]
    },
    {
      "id": "data",
      "name": "Data / Engineering",
      "desc": "Builds & runs data and AI systems",
      "deep_dive_sections": [
        "data_dd",
        "eng_dd"
      ]
    },
    {
      "id": "gov",
      "name": "Risk / Governance / Legal",
      "desc": "Risk, compliance, responsible AI",
      "deep_dive_sections": [
        "gov_dd"
      ]
    },
    {
      "id": "people",
      "name": "HR / Change / People",
      "desc": "Workforce, skills, adoption",
      "deep_dive_sections": [
        "people_dd"
      ]
    },
    {
      "id": "biz",
      "name": "Business / Product",
      "desc": "Operates a business unit or product",
      "deep_dive_sections": []
    }
  ],
  "sections": [
    {
      "id": "core",
      "title": "Core capabilities",
      "blurb": "Everyone answers these.",
      "optional": false,
      "questions": [
        {
          "id": "vision",
          "type": "scored_text",
          "category": "strategy",
          "name": "AI vision & ambition",
          "text": "How clear is your organisation's vision and ambition for AI? Is there a documented AI strategy, and who owns it?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "roadmap",
          "type": "scored_text",
          "category": "strategy",
          "name": "AI roadmap & investment",
          "text": "Is there a roadmap for AI with short- and long-term priorities? How are AI investments decided and sequenced?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "prioritise",
          "type": "scored_text",
          "category": "value",
          "name": "Use-case prioritisation",
          "text": "How do AI use cases get identified, evaluated and prioritised? Walk me through how a new idea becomes a funded project.",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "businesscase",
          "type": "scored_text",
          "category": "value",
          "name": "Business cases & value tracking",
          "text": "How do you build the business case for an AI initiative, and how is value actually measured after it ships?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "operatingmodel",
          "type": "scored_text",
          "category": "org",
          "name": "AI operating model",
          "text": "How is AI organised — central team, embedded, federated? Is there a defined operating model, or is it improvised per project?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "culture",
          "type": "scored_text",
          "category": "people",
          "name": "AI culture & change",
          "text": "How would you describe the organisation's appetite for AI? How is adoption and change managed?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "skills",
          "type": "scored_text",
          "category": "people",
          "name": "AI skills & gaps",
          "text": "How well do you understand the AI skills you have versus what you need? Is there a plan to close the gaps?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "training",
          "type": "scored_text",
          "category": "people",
          "name": "AI training & literacy",
          "text": "What AI training, upskilling or literacy programs exist, and who actually has access to them?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "ethics",
          "type": "scored_text",
          "category": "gov",
          "name": "Responsible AI policy",
          "text": "What policies or principles govern responsible and ethical use of AI? Are they actually enforced, or shelfware?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "risk",
          "type": "scored_text",
          "category": "gov",
          "name": "AI risk framework",
          "text": "How are AI-specific risks identified and managed? Is there a risk framework or register for AI?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "compliance",
          "type": "scored_text",
          "category": "gov",
          "name": "Compliance monitoring",
          "text": "How do you monitor and enforce compliance with AI policies and legal/regulatory obligations?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "architecture",
          "type": "scored_text",
          "category": "eng",
          "name": "AI architecture & platform",
          "text": "Describe the platforms and infrastructure you use to build and run AI. How standardised is it across teams?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "deploy",
          "type": "scored_text",
          "category": "eng",
          "name": "Deploy & operate AI",
          "text": "Once an AI solution is built, how does it get deployed, monitored and supported in production?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "dataacquire",
          "type": "scored_text",
          "category": "data",
          "name": "Acquire & prepare data",
          "text": "How easily can teams get the data they need for AI? Walk me through sourcing and preparing it.",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "dataquality",
          "type": "scored_text",
          "category": "data",
          "name": "Data quality & integrity",
          "text": "How confident are you in the quality, completeness and consistency of data used for AI?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "datagov",
          "type": "scored_text",
          "category": "data",
          "name": "Data governance for AI",
          "text": "What governance exists over data used in AI — ownership, access controls, lineage?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "datasec",
          "type": "scored_text",
          "category": "data",
          "name": "Security & privacy for AI data",
          "text": "How is sensitive data protected when used in AI? Any privacy controls specific to AI?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        }
      ]
    },
    {
      "id": "strategy_dd",
      "title": "Strategy deep-dive",
      "blurb": "Extra depth for the areas you know best.",
      "audience": [
        "exec"
      ],
      "optional": false,
      "questions": [
        {
          "id": "trendscan",
          "type": "scored_text",
          "category": "strategy",
          "name": "Trend scanning",
          "text": "How do you monitor and interpret emerging AI trends and assess their impact on your business?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "readiness",
          "type": "scored_text",
          "category": "strategy",
          "name": "Readiness measurement",
          "text": "Do you formally assess organisational readiness for AI? How?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "innovation",
          "type": "scored_text",
          "category": "value",
          "name": "Innovation & experimentation",
          "text": "How are new AI ideas and experiments fostered, captured and scaled?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "ecosystem",
          "type": "scored_text",
          "category": "org",
          "name": "Ecosystem & partnerships",
          "text": "How do you use external partners, vendors and alliances in your AI strategy?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        }
      ]
    },
    {
      "id": "data_dd",
      "title": "Data deep-dive",
      "blurb": "Extra depth for the areas you know best.",
      "audience": [
        "data"
      ],
      "optional": false,
      "questions": [
        {
          "id": "dataprep",
          "type": "scored_text",
          "category": "data",
          "name": "Data prep & feature engineering",
          "text": "How mature are your data preparation and feature-engineering practices for AI?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "dataintegration",
          "type": "scored_text",
          "category": "data",
          "name": "Data integration",
          "text": "How well integrated are your data sources for AI use, internal and external?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "lineage",
          "type": "scored_text",
          "category": "data",
          "name": "Data lineage & traceability",
          "text": "Can you trace the origin and transformation of data feeding AI models?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "metadata",
          "type": "scored_text",
          "category": "data",
          "name": "Metadata & semantics",
          "text": "What metadata, cataloguing or semantic-layer capabilities support AI?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "observability",
          "type": "scored_text",
          "category": "data",
          "name": "Data observability",
          "text": "How do you monitor the health and behaviour of data over time?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "analytics",
          "type": "scored_text",
          "category": "data",
          "name": "Analytics & visualisation",
          "text": "How are analytics and visualisation used to support AI insight?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        }
      ]
    },
    {
      "id": "eng_dd",
      "title": "Engineering deep-dive",
      "blurb": "Extra depth for the areas you know best.",
      "audience": [
        "data"
      ],
      "optional": false,
      "questions": [
        {
          "id": "buildvsbuy",
          "type": "scored_text",
          "category": "eng",
          "name": "Build vs buy",
          "text": "How do you decide build vs buy for AI capabilities? Is there a rule or is it case-by-case?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "buildapps",
          "type": "scored_text",
          "category": "eng",
          "name": "Building AI applications",
          "text": "Describe how AI applications and models actually get built and trained here.",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "testing",
          "type": "scored_text",
          "category": "eng",
          "name": "Testing AI systems",
          "text": "How are AI models and applications tested before and after release?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "techdebt",
          "type": "scored_text",
          "category": "eng",
          "name": "AI technical debt",
          "text": "How do you track and manage technical debt in AI systems?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "scale",
          "type": "scored_text",
          "category": "eng",
          "name": "Scaling AI systems",
          "text": "How do you scale AI systems that prove valuable?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "mlops",
          "type": "scored_text",
          "category": "eng",
          "name": "Operationalisation / MLOps",
          "text": "What MLOps practices keep AI running in line with service levels?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        }
      ]
    },
    {
      "id": "gov_dd",
      "title": "Governance deep-dive",
      "blurb": "Extra depth for the areas you know best.",
      "audience": [
        "gov"
      ],
      "optional": false,
      "questions": [
        {
          "id": "security",
          "type": "scored_text",
          "category": "gov",
          "name": "AI security & safety",
          "text": "What frameworks address AI-specific security and safety threats?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "legal",
          "type": "scored_text",
          "category": "gov",
          "name": "Legal & regulatory",
          "text": "How do you track and respond to AI legal and regulatory requirements?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "sustainable",
          "type": "scored_text",
          "category": "gov",
          "name": "Sustainable / cost governance",
          "text": "Are sustainability and cost-of-AI considerations governed at all?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "govteams",
          "type": "scored_text",
          "category": "gov",
          "name": "Governance teams & decision rights",
          "text": "Who governs AI? Are there cross-functional teams and clear decision rights?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "oversight",
          "type": "scored_text",
          "category": "gov",
          "name": "Oversight & assurance",
          "text": "What systems provide continuous oversight and assurance of AI in production?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        }
      ]
    },
    {
      "id": "people_dd",
      "title": "People deep-dive",
      "blurb": "Extra depth for the areas you know best.",
      "audience": [
        "people"
      ],
      "optional": false,
      "questions": [
        {
          "id": "workforce",
          "type": "scored_text",
          "category": "people",
          "name": "Workforce planning",
          "text": "How do you forecast and plan for AI's impact on roles and staffing?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "recruit",
          "type": "scored_text",
          "category": "people",
          "name": "AI talent acquisition",
          "text": "How do you attract and recruit AI talent?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "trainingeff",
          "type": "scored_text",
          "category": "people",
          "name": "Training effectiveness",
          "text": "How do you measure whether AI training actually changes anything?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        },
        {
          "id": "changemgmt",
          "type": "scored_text",
          "category": "people",
          "name": "Change management",
          "text": "How are AI-driven change initiatives executed and monitored?",
          "scale": "maturity5",
          "scored": true,
          "star": false,
          "ai_drafted": false
        }
      ]
    }
  ],
  "grids": {},
  "outputs": [
    "Overall maturity picture + by-area",
    "Contested-capability view",
    "Coverage check",
    "Confidence ramp",
    "Executive summary + priority gaps",
    "Printable client report"
  ],
  "report_defaults": [
    "rpt-maturity-standard"
  ],
  "playbook": {
    "sequence": [
      "strategy",
      "gov",
      "data",
      "org",
      "people",
      "eng",
      "value"
    ],
    "sequence_note": "Foundations first: strategy and governance decisions shape everything downstream; data and operating model make delivery repeatable; engineering and value work then compound instead of stalling.",
    "actions": {
      "strategy": {
        "to_2": [
          "Name an executive owner for AI and write down the ambition in one page: where AI matters to the business and where it does not.",
          "Inventory what is already happening with AI across the organisation, sanctioned or not."
        ],
        "to_3": [
          "Publish an AI strategy tied to business goals, with a 12–24 month roadmap and a decision forum that meets.",
          "Agree the handful of priority domains and say no to the rest for now."
        ],
        "to_4": [
          "Set measurable targets per priority domain and review progress against them each quarter.",
          "Fund the roadmap as a portfolio, not project by project."
        ],
        "to_5": [
          "Refresh the strategy on a cadence driven by results and market shifts, with scenario planning for model and regulatory change."
        ]
      },
      "value": {
        "to_2": [
          "Keep one shared list of AI use cases with an owner and a one-line value hypothesis each.",
          "Stop funding ideas that bypass the list."
        ],
        "to_3": [
          "Adopt a single intake and prioritisation method (value × feasibility × risk) and apply it to every candidate.",
          "Define what 'done' means for a pilot: a decision to scale, hold or stop, with the evidence required."
        ],
        "to_4": [
          "Track realised value against the hypothesis for every scaled use case and report it alongside cost.",
          "Retire or rework use cases that are not paying back."
        ],
        "to_5": [
          "Run the portfolio as a continuous funnel with stage gates, benchmarks and reinvestment of returns."
        ]
      },
      "org": {
        "to_2": [
          "Assign clear accountability for AI delivery, even if it sits in an existing team for now.",
          "Write down how an AI idea gets resourced today so the gaps are visible."
        ],
        "to_3": [
          "Choose an operating model (central, hub-and-spoke or federated) and staff the core roles: product, data, engineering, risk.",
          "Define the interfaces with IT, data and the business units."
        ],
        "to_4": [
          "Measure delivery throughput and cycle time and remove the recurring blockers.",
          "Formalise a community of practice that shares patterns and reusable components."
        ],
        "to_5": [
          "Let the model flex by domain with shared platforms and standards, and review it against outcomes annually."
        ]
      },
      "people": {
        "to_2": [
          "Run a skills stocktake: who can do what today, and who wants to learn.",
          "Give leaders a shared, plain-language briefing on what AI can and cannot do."
        ],
        "to_3": [
          "Put an AI literacy programme in place by role type, with a small number of deep specialists.",
          "Set expectations for how staff should and should not use AI tools, and support the change explicitly."
        ],
        "to_4": [
          "Tie AI skills to career paths and performance goals; measure adoption and confidence, not just attendance.",
          "Build change management into every AI rollout by default."
        ],
        "to_5": [
          "Sustain a learning culture where teams experiment safely and share what they learn across the organisation."
        ]
      },
      "gov": {
        "to_2": [
          "Adopt a short, enforceable responsible-AI policy covering acceptable use, data and human oversight.",
          "Create a register of AI systems in use and who owns each."
        ],
        "to_3": [
          "Stand up an AI risk assessment that every use case passes before build and before scale, proportionate to impact.",
          "Assign a governance forum with the authority to stop a deployment."
        ],
        "to_4": [
          "Monitor deployed systems for drift, bias and incidents, with defined response paths.",
          "Audit against the policy and the emerging regulatory obligations at least annually."
        ],
        "to_5": [
          "Embed governance into the delivery platform so compliance is largely automatic, and adapt controls as regulation and models change."
        ]
      },
      "eng": {
        "to_2": [
          "Pick one sanctioned path to build and run an AI solution, even if minimal, and route new work through it.",
          "Get the current pilots under source control and basic monitoring."
        ],
        "to_3": [
          "Standardise the platform: environments, model access, deployment pipeline, logging and cost controls.",
          "Define engineering standards for prompts, evaluation and testing of AI components."
        ],
        "to_4": [
          "Instrument quality, latency, cost and usage for every production system and act on the trends.",
          "Reuse components (retrieval, guardrails, evaluation) across use cases rather than rebuilding."
        ],
        "to_5": [
          "Run a self-service platform with automated evaluation and rollback, evolving as model capabilities shift."
        ]
      },
      "data": {
        "to_2": [
          "Identify the data the priority use cases depend on and who owns it.",
          "Fix the most damaging quality issues in that data first, not everywhere."
        ],
        "to_3": [
          "Put data governance in place for AI workloads: ownership, quality rules, classification, access controls and lineage.",
          "Make the priority data discoverable and reliably available to delivery teams."
        ],
        "to_4": [
          "Measure data quality and availability as a service level and report against it.",
          "Extend governance to unstructured content and model inputs and outputs."
        ],
        "to_5": [
          "Treat data as a managed product with automated quality, privacy-preserving access and continuous fitness-for-AI checks."
        ]
      }
    }
  }
}