AI Readiness Assessment Glossary
The complete reference for Recser's 39 AI readiness criteria across 10 dimensions. Each criterion includes four scored maturity-level options — from Explorer (0 pts) to Visionary (3 pts) — covering Strategy, Data, Talent, AI Governance, Culture, Deployment, Business Impact, Scaling, and Ecosystem.
About This Glossary
This is the complete reference for the Recser AI Readiness Assessment — Canon V4.2, 533A Panel Approved. The assessment covers 39 criteria across 10 dimensions. Each criterion is scored 0–3 points corresponding to four maturity levels: Explorer (0), Builder (1), Scaler (2), and Visionary (3).
Module 1: Strategy and Vision — Where are we going with AI?
Evaluates if you have a clear, owned, and funded plan for AI adoption.
- Documentation: Which statement best describes your organisation's AI strategy?
- (0 pts) No written AI strategy — Conversations are informal and occasional
- (1 pt) Emerging AI roadmap — Exists but not formally approved or integrated into business plans
- (2 pts) Formal AI strategy — Approved by leadership and aligned with key business goals
- (3 pts) AI-Centric Business Strategy — Every major decision considers AI implications
- Ownership: Who owns AI strategy in your organisation?
- (0 pts) No Clear Owner — AI is discussed informally; no one is accountable
- (1 pt) IT or Mid-level Manager — AI treated as technical project, not business priority
- (2 pts) Senior Executive (C-Suite) — Specific leader has budget and accountability for AI success
- (3 pts) CEO or Board Level — Strategic priority integrated into the entire business
- Pipeline: How developed is your pipeline of AI projects?
- (0 pts) Brainstorming Ideas — No active projects or clear value defined yet
- (1 pt) Running Pilots — Testing initial use cases to prove value
- (2 pts) Use Cases in Production — Several solutions deployed and delivering measurable value
- (3 pts) Continuous Innovation Pipeline — AI embedded across business functions with systematic scaling
- Objectives: What are your primary AI goals?
- (0 pts) Learning & Exploration — Understand basics and identify potential applications
- (1 pt) Proving Value — Run pilots to demonstrate feasibility and ROI
- (2 pts) Operational Efficiency — Improve speed, cost, and quality across the organisation
- (3 pts) Business Transformation — Create new business models, products, or markets
Module 2: Data Readiness — Do we have the data foundation?
Checks if your data is clean and available enough to fuel AI.
- Quality: Which statement best describes the quality and reliability of your data?
- (0 pts) Fragmented & Inconsistent — Data trapped in silos with frequent errors and gaps
- (1 pt) Usable but Manual — Reliable data requires significant manual cleaning before use
- (2 pts) High Quality & Trusted — Centralized data with automated validation and team trust
- (3 pts) Real-time Precision — Continuously monitored, self-correcting data with industry-leading accuracy
- Access: How quickly can teams access the data they need for AI projects?
- (0 pts) Days or Weeks — Manual IT requests; data is trapped in silos
- (1 pt) Within 24 Hours — Partial self-service; teams view reports but need data help
- (2 pts) On-Demand (Minutes) — Unified platform allows authorised users immediate access
- (3 pts) Instant & Automated — Direct APIs allow AI systems to fetch real-time data
- Governance: Which statement best describes your data governance maturity?
- (0 pts) Ad-Hoc & Informal — No formal policies; ownership unclear and managed reactively
- (1 pt) Basic Policies — Drafts exist; key stewards appointed for main datasets
- (2 pts) Defined Framework — Comprehensive rules; ownership, privacy, and security standards enforced
- (3 pts) Automated & Embedded — Governance baked into platform; rules enforced automatically
- Infrastructure: How mature is your data infrastructure (storage and pipelines)?
- (0 pts) Manual & Siloed — Manual preparation in spreadsheets; high effort to access data
- (1 pt) Semi-Automated — Basic scripts move data, but pipelines are fragile and break
- (2 pts) Scalable Data Platform — Centralized warehouse with reliable, fully automated DataOps pipelines
- (3 pts) Real-Time & AI-Native — Infrastructure supports streaming data and dedicated Feature Stores
Module 3: Technology and Infrastructure — Do we have the tools?
Determines if you have the compute power to run modern AI.
- Deployment: What is the level of sophistication of your AI deployment?
- (0 pts) Individual Ad-hoc Use — Staff use tools like ChatGPT individually without integration
- (1 pt) Standard Integrations — Using AI features built into existing software with minimal configuration
- (2 pts) Customized Context — AI models connected to company data to perform specific tasks
- (3 pts) Autonomous Agents — AI acts autonomously to execute complex workflows across multiple systems
- Models: How do you manage the lifecycle of your AI models (versioning, deployment, and monitoring)?
- (0 pts) Ad-hoc & Manual — Models stored locally with no version history or tracking
- (1 pt) Basic Versioning — Code tracked in Git but deployment remains a manual process
- (2 pts) Automated Pipelines — Fully automated deployment with monitoring and rollback capabilities
- (3 pts) Continuous Operations — Self-optimizing loops that automatically retrain based on new data
- Compute: What computing resources support your AI workloads?
- (0 pts) Standard Hardware — No specialised chips; runs on standard laptops or servers
- (1 pt) On-Demand Cloud — Rent access to cloud GPUs or models when needed
- (2 pts) Scalable Infrastructure — Automated clusters scale up for heavy training or inference
- (3 pts) Orchestrated Ecosystem — Workloads automatically route to the most efficient hardware
- Integration: How integrated are AI capabilities with your existing business systems?
- (0 pts) Isolated & Standalone — Tools used separately with no connection to internal systems
- (1 pt) Point-to-Point Connections — Direct links exist for specific use cases but are fragile
- (2 pts) Integrated Platform — AI built into core applications allowing smooth data flow
- (3 pts) Composable Ecosystem — Modular layer connecting everything; models swapped without breaking systems
Module 4: Talent and Skills — Do we have the people?
Assesses if your people have the skills to execute AI initiatives.
- Specialists: How is your AI talent organized and resourced?
- (0 pts) Outsourced or None — Rely entirely on external vendors; no dedicated internal staff
- (1 pt) Centralised Core Team — Small central team handles all AI requests for the organisation
- (2 pts) Embedded Experts — Specialists sit within business units to drive specific goals
- (3 pts) World-Class Talent — Top-tier global talent driving proprietary innovation
- Literacy: How widespread is AI literacy across your organisation?
- (0 pts) Limited to Tech Teams — Only IT specialists understand AI; others are unaware
- (1 pt) Role-Specific Training — Upskilling limited to specific technical or analytical roles
- (2 pts) Company-Wide Foundations — AI literacy included in onboarding; staff understand basics
- (3 pts) Universal Fluency — AI is core competency; staff use tools daily
- Development: Which statement describes your AI talent development?
- (0 pts) Self-Driven Only — Staff learn on their own time using free resources
- (1 pt) Ad-Hoc Training — Occasional workshops or subscriptions for interested individuals
- (2 pts) Structured Pathways — Formal certifications and clear career tracks for AI roles
- (3 pts) Innovation Culture — Hackathons, R&D time, and knowledge sharing are standard
- Collaboration: How do technical and business teams collaborate on AI?
- (0 pts) Siloed (IT-Led) — Tech builds in isolation; business units are passive customers
- (1 pt) Consultative Approach — Tech and Business collaborate on projects but remain separate
- (2 pts) Integrated Squads — Cross-functional teams sit and work together permanently
- (3 pts) Federated Model — Business units lead initiatives, supported by central platform team
- Retention: How successful are you at attracting and retaining top AI talent?
- (0 pts) Struggling to Hire — Struggle to recruit qualified staff; high turnover or contractor reliance
- (1 pt) Vendor Dependent — Rely on external partners for complex or specialised AI work
- (2 pts) Competitive Employer — Successfully recruit and retain experienced AI engineers and scientists
- (3 pts) Talent Magnet — Destination of choice; top-tier talent proactively seeks us out
Module 5: AI Governance — Do we have the rules?
Measures if you are building solutions safely, ethically, and legally.
- Ethics: How are AI ethics and safety principles applied in your organisation?
- (0 pts) No Formal Policy — No documented principles; decisions left to individual judgment
- (1 pt) Guiding Principles — High-level values written down but not strictly enforced
- (2 pts) Mandatory Review — Ethics checklist or committee review required before launch
- (3 pts) Ethical by Design — Safety and fairness checks built into development process
- Oversight: How do you ensure humans review and approve AI decisions?
- (0 pts) No Formal Oversight — Rely on individual users; no mandatory review process
- (1 pt) Ad-Hoc Review — Humans intervene only for high-risk or suspicious decisions
- (2 pts) Mandatory Human-in-the-Loop — Key decisions automatically pause for human approval before proceeding
- (3 pts) Independent Audit — Independent teams regularly audit effectiveness of human reviews
- Fairness: How do you monitor AI systems for bias and fairness?
- (0 pts) No Monitoring — We do not check for bias; assume outputs are objective
- (1 pt) Reactive Checks — Manual audits performed only when users complain or issues suspected
- (2 pts) Continuous Auditing — Automated tools regularly scan outputs for bias and compliance
- (3 pts) Real-Time Mitigation — System automatically detects and blocks biased outputs in real-time
- Compliance: How mature is your AI risk management and regulatory compliance?
- (0 pts) Unmanaged Exposure — No process to identify or mitigate AI regulatory risks
- (1 pt) Ad-Hoc Assessments — Manual risk checks for specific projects; consistency varies
- (2 pts) Systematic Framework — Standard framework followed with clear incident response plans
- (3 pts) Continuous Compliance — Automated checks aligned with global standards like ISO 42001
Module 6: Organisational Culture — Is our organisation ready to change?
Tests your team's willingness to experiment, fail, learn, and adapt.
- Sentiment: What is the predominant employee attitude toward AI?
- (0 pts) Fear & Resistance — High anxiety about job loss; active resistance to tools
- (1 pt) Cautious Curiosity — Interest exists, but adoption is slow due to uncertainty
- (2 pts) Proactive Experimentation — Teams willingly try tools; failure seen as learning
- (3 pts) Innovation DNA — AI culturally embraced; staff actively seek automation opportunities
- Leadership: Which leadership style best characterises your organisation's approach to innovation?
- (0 pts) Command & Control — Top-down decisions; innovation stifled by strict hierarchy
- (1 pt) Risk-Averse Management — Focus on stability; pilots permitted but hesitation to scale
- (2 pts) Collaborative Agile — Leaders actively encourage cross-team feedback and rapid iteration
- (3 pts) Transformational Leadership — Leaders remove roadblocks and empower teams to take risks
- Failure: How does your organisation handle AI experiments that fail?
- (0 pts) Failure Punished — Failed projects damage careers; teams avoid risk to protect reputation
- (1 pt) Failure Tolerated — Failures accepted if cheap, but rarely discussed openly
- (2 pts) Structured Reviews — Mandatory post-mortems; lessons documented to prevent repeating mistakes
- (3 pts) Fail-Fast Culture — Failure expected and budgeted for; insights applied to next sprint
- Workflows: How deeply is AI embedded into daily roles and performance goals?
- (0 pts) Not Defined — AI absent from job descriptions; usage is informal
- (1 pt) Specialist Focus — Responsibilities limited to data and tech teams
- (2 pts) Standard Enabler — Teams use tools; proficiency tracked in key roles
- (3 pts) Reinvented Roles — Jobs redesigned; staff rewarded for automating workflows
Module 7: Strategic Deployment — How are we deploying AI?
Assesses how you effectively delegate to AI without losing control.
- Selection: How do you decide which tasks are safe and suitable for AI automation?
- (0 pts) Gut Feeling — No formal method; choices based on excitement or need
- (1 pt) Basic Feasibility — Check technical capability but without deep risk analysis
- (2 pts) Risk vs Value Mapping — Categorise tasks by cost of error and potential value
- (3 pts) Knowledge-Structure Matrix — Separate repetitive, rule-based tasks from those needing human judgment
- Oversight: How do you determine the level of human oversight for different AI applications?
- (0 pts) One-Size-Fits-All — Same testing and oversight applied regardless of risk
- (1 pt) Ad-Hoc Judgment — Supervision levels decided by leads based on intuition
- (2 pts) Risk-Based Tiers — Tools strictly classified based on their risk level
- (3 pts) Adaptive Autonomy — System runs autonomously but flags humans when needed
- Transparency: How much visibility do you have into why an AI system made a specific decision?
- (0 pts) Black Box — Input and output visible; internal connection logic remains unknown
- (1 pt) Behavioral Testing — Rigorous testing maps likely system behaviour without internal visibility
- (2 pts) Partial Visibility — Top factors influencing results are identified and understood
- (3 pts) Auditable Reasoning — System provides citations or chain-of-thought for full logic audit
Module 8: Business Impact — What value are we creating?
Tracks the real financial and mission value you are creating with AI.
- Costs: How do you manage and optimize the running costs of your AI systems?
- (0 pts) Unmeasured Costs — No visibility; costs buried in general IT budgets
- (1 pt) Total Cost Monitoring — Track total monthly bill but cannot break it down further
- (2 pts) Attributed Costs — Accurately allocate costs to specific projects or departments
- (3 pts) Unit Economics Optimised — Measure cost per transaction and actively optimise for sustainable scaling
- ROI: How do you measure the return on investment or mission value of your AI projects?
- (0 pts) No Formal Tracking — Rely on anecdotes; value is assumed but not calculated
- (1 pt) Technical Metrics Only — Track model performance but not business or mission outcomes
- (2 pts) Efficiency Gains — Track resource savings like reduced costs or time saved
- (3 pts) Strategic Value Realisation — Measure returns against revenue, social impact, or core goals
- Speed: On average, how long does it take to move an AI concept into production?
- (0 pts) No Production Deployment — Have not yet successfully deployed an AI model to production
- (1 pt) Long Cycles (12+ Months) — Slow manual deployment; projects often stuck in pilot phase
- (2 pts) Standard Cycles (6–12 Months) — Reliable process, but scaling requires significant custom effort
- (3 pts) Rapid Agility (< 6 Months) — Automated pipelines allow us to deploy value quickly
- Maturity: What is the highest level of value AI has actively delivered to date?
- (0 pts) No Proven Value — Experimenting or piloting; no measurable results observed yet
- (1 pt) Operational Efficiency — Automating existing tasks to save time or reduce costs
- (2 pts) Product & Service Enhancement — Improving quality or user experience of existing offerings
- (3 pts) Business Transformation — Creating new revenue streams or new mission delivery models
Module 9: Scaling AI — How do we grow AI initiatives?
Evaluates your ability to move from pilot to mass production.
- Expansion: How do you expand successful AI pilots to the rest of the organisation?
- (0 pts) Stuck in Pilot — Pilots remain isolated; no plan for wider adoption
- (1 pt) Ad-Hoc Expansion — Scaling is manual; relies on leaders pushing case-by-case
- (2 pts) Standardised Playbook — Documented process to roll out successful tools organisation-wide
- (3 pts) Industrialised AI Factory — Shared infrastructure makes scaling rapid, repeatable, and automated
- Platform: How are your AI solutions architected to support scaling and integration?
- (0 pts) Isolated Silos — Standalone experiments with no reusable code or architecture
- (1 pt) Centralised Hosting — Common environment used; integration remains difficult and manual
- (2 pts) API-First Design — Models built as APIs to plug into existing software
- (3 pts) Universal Deployment — Containerised solutions run seamlessly across any environment
- Resources: How do you ensure long-term funding and resources for AI scaling?
- (0 pts) No Dedicated Budget — Resources are hunted project-by-project without long-term security
- (1 pt) Pilot-Only Funding — Budget exists for pilots but not for long-term maintenance
- (2 pts) Dedicated Allocation — Recurring budget or grant specifically allocated to support AI scaling
- (3 pts) Sustainable Economics — Value generated effectively covers the cost of ongoing operations
- Assurance: How do you maintain high performance as you scale AI across the organisation?
- (0 pts) Reactive Fixes — Fix problems only on user complaint; no proactive testing
- (1 pt) Manual Quality Checks — Occasional human spot-checks; quality dips as volume increases
- (2 pts) Standardised QA Protocols — Strict testing rules ensure consistency across all teams
- (3 pts) Automated Retraining Loop — Systems detect drift and trigger retraining without manual intervention
Module 10: Ecosystem and Partnerships — Who do we work with?
Assesses the strength of your network of vendors and experts.
- Vendors: How does your organisation engage with external AI technology providers?
- (0 pts) Ad-Hoc Procurement — Buy off-the-shelf tools as needed with no deeper relationship
- (1 pt) Managed Relationships — Selected key vendors; interactions limited to support and licensing
- (2 pts) Strategic Integration — Work closely with vendors to customise tools for specific workflows
- (3 pts) Co-Innovation Ecosystem — Actively co-develop solutions creating new IP and shared capabilities
- Research: How deeply do you partner with external research bodies on AI initiatives?
- (0 pts) No Research Engagement — No engagement with research institutions regarding AI
- (1 pt) Passive Consumer — Attend conferences and read reports; no active collaboration
- (2 pts) Project-Based Experimentation — Provide data or problems for student projects or pilots
- (3 pts) Joint R&D Partnership — Formally co-develop new models, IP, or impact studies
- Sharing: How do you share knowledge and learn from industry peers regarding AI?
- (0 pts) Internal Focus — Focus on internal projects; no active external engagement
- (1 pt) Passive Participant — Attend webinars and read reports; rarely contribute insights
- (2 pts) Active Contributor — Regularly share learnings at conferences and working groups
- (3 pts) Ecosystem Leader — Actively define industry standards and frameworks others adopt
Maturity Levels
Each criterion is scored on a four-level scale. The four levels are: L1 Explorer (0-25%, early stage capabilities), L2 Builder (26-50%, developing foundational capabilities), L3 Scaler (51-75%, established and expanding capabilities), and L4 Visionary (76-100%, industry-leading capabilities). Scores are aligned with ISO/IEC 42001 and the NIST AI Risk Management Framework.