AI Fluency Matrix

Core human competencies for an automation-first organisation

A new reality for work

Foundational expectations for every role

The AI Fluency Matrix describes the base competencies expected of an AI first colleague.

AI is becoming part of the infrastructure. Our advantage comes from how deeply we work with it each day. This matrix sets expectations for daily, meaningful AI use in core workflows, not occasional demos or shallow tasks, and highlights the human skills that grow in value as agents and automation spread.

Levels are indicative. Most roles will blend Foundation and Proficient, with some specialists operating mainly at the Advanced level for selected competencies. AI fluency is expected at least at the Foundation level for everyone.

CompetencyDescriptionFoundationProficientAdvanced
Clarity and IntentDefining the “why” and making deliberate choices about what should exist. The core craft when generation is cheap and intermediary artifacts are reduced.
  • Can articulate the basic problem and goal before using AI to explore solutions.
  • Asks “what are we trying to achieve?” to frame AI-assisted work.
  • Recognises when an AI output, while plausible, doesn't serve the core purpose.
  • Uses AI as a thinking partner to sharpen and challenge the definition of the problem.
  • Clearly communicates intent to stakeholders and in prompts, focusing on outcomes over assets.
  • Sequences work based on clarity of intent, prioritising what needs human reasoning first.
  • Shapes meaning, coherence, and direction for projects and teams, not just outputs.
  • Sets a clear “why” that guides AI system behaviour and agent workflows, ensuring alignment with human values.
  • Elevates taste and discernment as a creative act, defining what good looks like for novel challenges.
Critical thinkingAbility to question assumptions, evaluate evidence, and spot plausible nonsense in AI outputs before it reaches customers or stakeholders.
  • Checks AI outputs against basic facts, constraints, and source material.
  • Notices obvious contradictions, gaps, or overconfident answers.
  • Asks follow up questions when something does not feel right rather than accepting it.
  • Applies simple tests such as source checks, comparison across models, and sanity checks.
  • Identifies subtle bias, missing context, or shallow reasoning in AI answers.
  • Articulates why a given output should be accepted, revised, or rejected, and captures that learning.
  • Designs review patterns that reliably catch known failure modes in AI assisted workflows.
  • Coaches others in how to challenge AI outputs without losing momentum.
  • Builds guardrails and evaluation loops that reduce the chance of poor decisions at scale.
Systems thinkingUnderstanding how data, tools, people, and processes connect. Shaping system behaviour, responses, and flows to curate an intended experience.
  • Can describe the basic flow of work from input to AI steps to human review and output.
  • Recognises when one step depends on another or on a specific data source.
  • Flags when a change to an AI workflow might affect other teams or systems.
  • Maps end to end flows, including prompts, tools, data, and decision points.
  • Anticipates unintended consequences when new automation is added or an agent takes on more work.
  • Identifies leverage points where a small change improves the whole system, not just one step.
  • Designs AI enabled systems that are observable, resilient, and adaptable as models evolve.
  • Uses feedback loops and data exhaust to improve quality, speed, and safety over time.
  • Helps the organisation think in loops and workflows, not one off tools or pilots.
Judgement and TasteKnowing what good looks like for your domain, and choosing the option that best fits context and purpose. Producing clarity of intent is a key creative act.
  • Can explain in plain language why one AI output is better than another.
  • Aligns choices with existing guidelines, examples, and known constraints.
  • Avoids obviously off brand, generic, or low quality outputs, even when they are fast.
  • Tunes AI prompts and workflows to produce outputs that closely match the desired style and intent.
  • Balances speed with quality so work ships on time without feeling commoditised.
  • Actively curates examples of good work and shares them as references for others and for agents.
  • Sets and evolves the standard for what good looks like in a function or team.
  • Maintains coherence and meaning across multiple agents, touchpoints, and channels.
  • Uses narrative and discernment to lift the overall bar as AI takes on more production work.
Creative Problem SolvingUsing non linear thinking and experimentation to navigate novel or messy problems, with AI as a collaborator and thinking partner rather than a vending machine.
  • Tries more than one prompt, model, or tool when the first attempt fails.
  • Breaks down vague problems into smaller, clearer questions for AI and for people.
  • Uses AI to quickly explore multiple directions without premature commitment.
  • Uses AI as a thought partner to explore options, combine perspectives, and bring clarity to a problem space.
  • Reframes problems from different angles to uncover better paths or beachheads.
  • Builds quick prototypes, drafts, or simulations to test ideas with low risk and high learning.
  • Designs structured explorations for complex, ambiguous challenges that combine humans and agents.
  • Helps teams escape local optima by challenging the problem framing and suggesting new use cases.
  • Uses constraints as a tool to focus creativity and explore the intent behind choices.
AI Fluency and CollaborationWorking with AI as a colleague in daily, meaningful work. Knowing when to hand work to machines, when to intervene, and where deep human judgement is non negotiable.
  • Uses AI several times a week in real work, beyond generic email drafting.
  • Can name at least two core workflows where AI or an agent meaningfully assists their role.
  • Understands the basic limits of the models they use and checks outputs before acting.
  • Works more conversationally with AI, using it daily in core workflows for acceleration and thinking.
  • Maintains a small personal prompt or pattern library for recurring tasks.
  • Designs simple agent or tool chains to support their role and shares them with others.
  • Re-architects workflows around agents and automation where it gives compounding advantage.
  • Monitors, validates, and improves digital teammates, treating agent operations as part of the job.
  • Drives deep, sticky AI adoption in their area, focusing on daily use in the work that matters most.

Using the matrix

This matrix is a starting point for conversations about expectations, growth, and hiring in an AI enabled organisation. It can be used for role design, personal development plans, or as a shared language for what it means to be an AI first colleague who works with agents every day.

Over time, Ortomate will add role specific overlays, adoption targets, and assessment tools to help teams measure and grow these competencies in a calm, deliberate way, with daily deep AI use as a core leading indicator of capability.