Permission to question
People can challenge an AI recommendation, seek expertise and raise concerns without being penalized for slowing a decision.
Loading the Directing Intelligence competency model…
The Directing Intelligence competency model
AI changes what people can do. Human Skills shape what they choose to do, how well they do it, and what they take responsibility for.
Explore the eight levels
Directing Intelligence means directing AI toward valuable outcomes, elevating its work and remaining accountable. Explore what that requires at each level.
Our workforce recommendation: build AI Agility in everyone. Go further when the work requires it. L2 equips people to reshape human-led work. The levels beyond it address different workflow, system, leadership and innovation responsibilities.
Find a pathway →Choose a role or describe what someone needs to do. Explore a suggested destination and staged learning milestones.
Get a recommendation →I know the workDescribe the work, choose a few learning priorities and take away a development plan you can discuss and use.
Build a profile →Human Skills build agency
Look for the person’s understanding, questions, revisions and decisions. Those contributions show how Human Skills support the ability to choose, judge and act.
See what agency looks like in practice →A practical starting point
Describe the outcome, the people affected and the decisions someone must own. Then use the model to identify the capability the work requires.
Find a development pathway →A practical workforce development plan
Choose a role or tell us what someone needs to be able to do. Get a proposed destination, focused learning priorities and a timeline from your start date.
Already know what the role needs? Build your own competency profile →
The competency profiles
Select a level to see the work, human contribution, responsible practice and evidence that define it. Use the profile as a starting point for a role or learning pathway.
Build your own profile
Tell us what someone needs to do. Choose a few learning priorities, then build a plan for practice and review.
The complete competency matrix
Twenty capabilities. Eight levels. Four views of the same performance. Select a level heading for a complete profile, or a cell for the competency in detail.
Read across for progression. Read down for a level. Open a cell for the complete competency.
Human capability & agency
Human Skills enable people to understand, choose, create, question and act. The model connects those skills to specific human contributions, with observable behaviors from the taxonomy.
A connected model
Agency is expressed when a person uses those capabilities to direct the work and retains meaningful choice, judgment and responsibility. It also depends on conditions: authority, time, access, support and the ability to challenge or stop.
Assess the contribution in context. Naming a skill or producing an AI-assisted artifact is not evidence by itself that agency has grown.
Look for the human contribution
People can challenge an AI recommendation, seek expertise and raise concerns without being penalized for slowing a decision.
People have relevant knowledge, usable evidence, time and practice. A review gate is meaningful only when someone can perform the review.
Decision rights, intervention routes and accountability are explicit. A named human owner can actually change or stop the work.
Research, evidence & responsible use
Use the framework to build capability through real work. Understand where the source evidence ends and the proposed curriculum design begins.
Ten primary references inform the collaboration capabilities, critical evaluation, workforce use and responsible practice.
Capability Domains, Human Skills and Micro-Skills preserve the source workbook’s exact labels, behaviors and parent relationships.
Use authentic work, a reasoned human contribution and meaningful variation. Calibrate judgments for the context and consequence.
The level descriptors, curriculum mappings, role examples and development plans are proposed applications. L2 for everyone is HumanSkills’ workforce-development recommendation. Suggested dates are planning milestones, not externally validated deadlines. Contextualize them with the people doing the work, subject-matter experts and affected stakeholders.
Role presets suggest responsibilities to confirm. Expected-work suggestions use transparent matching rules and show the statements behind them. An unfamiliar or ambiguous requirement needs human interpretation. The tools do not diagnose a person or generate a readiness score.
The original editorial images were created with AI for this experience. They illustrate human collaboration and do not depict named learners, clients or program participants.
Role presets describe possible work, then ask the visitor to confirm the responsibility. Expected-work suggestions use a transparent set of matching rules. The result shows the supporting statements, any unresolved interpretation and editable targets.
Learning windows are HumanSkills planning recommendations. The research below supports contextual role profiles and differentiated learning needs; it does not establish our level mappings or durations.
