H3 Design OS
A Design Practitioner’s Map and Instrument
A three-layer framework for the whole practitioner: the H3 model, the π-shaped career architecture, and an operational guide for running design at scale. Twenty-five years of practice, made into a system.
For the designer who has the work but not the words, the practitioner at a crossroads, and the leader building what they wish had existed when they began.
An idea two hundred
and fifty years old.
Before this framework, there was a Swiss schoolteacher who believed that education was failing human beings because it taught only one part of a person at a time.
Johann Heinrich Pestalozzi (1746–1827) spent his life building schools for the poorest children in Europe. His conviction was simple and radical: that a whole human being thinks, feels, and acts, and that any system of development which attends to only one produces a diminished person. His guiding motto was “Learning by head, heart and hand.”
Thinking, reasoning, the capacity to understand the world through logic and observation.
Values, empathy, the capacity to feel the world through others and to act with conscience.
Skills, making, the capacity to act on the world: to build, create, and produce with craft.
The idea did not stay with Pestalozzi. It travelled (through Dewey, Steiner’s Waldorf, the Bauhaus, Outward Bound, and 4-H) and returns in contemporary management writing under different labels. H3 does not invent the triad. It inherits a 250-year-old formulation and applies it to a problem it has not yet been applied to: the whole practitioner in the design profession. The longer arc → Provenance.
Pestalozzi began with Head: intellect first, because he was teaching children how to know the world. H3 reorders to Hand → Head → Heart, because a designer’s accountability expands outward from a different starting point: first to the work, then to the path, then to the people.
The reordering is not stylistic. Each H sits at a different layer because each one becomes the bottleneck at a different stage: Hand when you are learning to make, Head when you are choosing where to specialise, Heart when others begin to depend on you.
This is not a moral preference either. Across twenty-five years of practice and hundreds of designers, the pattern holds: most enter through Hand. A few rare ones enter through Head, because they come from research or strategy backgrounds; a few through Heart, because they come from service design or therapeutic practice. A framework for development has to start where development actually starts.
“Pestalozzi asked what a child needs to become whole. H3 asks what a designer needs to become whole, and what a design organisation needs to stay that way. The triad is the same. The application, the order, and the unit of analysis are new.”
Back in 1997, a Software Engineering lecturer mentioned, almost in passing, that the deepest learning moves through head, heart, and hand, then continued with the lesson. I have not fully moved on from that aside. Some ideas arrive quietly and take root for a lifetime. This framework is, among many things, my almost-thirty years of gratitude for a moment that lasted no more than a breath.
Three layers.
One operating system.
H3 (Hand, Head, Heart as Design Operating System) is not a single framework. It is a system of three interlocking layers that answer three different questions about design practice.
For the designer who has the work but not the words
✦ Hand — your relationship with the work
The Model is where craft becomes legible. By the end of this layer you will have named your dominant H, chosen upstream or downstream, and picked one subject to study this week, with a definition of what done looks like.
For the stalled lead, and the designer at the Branch
◈ Head — your relationship with the path
The Structure is where you choose your direction deliberately. By the end of this layer you will have an honest read on whether you have plateaued under your current shape, whether the lead track or the deep-IC track fits your real energy, and one move to make this month.
For the leader running a team
♥ Heart — your relationship with the people
The Conduct is where accountability shifts to the ensemble. By the end of this layer you will have scored yourself across the five management domains, chosen one as this quarter's focus, and named the first ritual you will change.
“Each layer answers a different question. Together, they answer the one question that matters: what does it take to build and sustain excellent design practice over time?”
The thinking
behind H3
Practitioner-centric
H3 does not evaluate ideas; it develops the person who makes them. The question at the centre is not “is this viable?” but “can I engage with all three forces?” Designer growth is the unit of measurement.
Career-level integrated
H3 maps directly onto the 6-level career architecture. L1–L2 designers are commonly single-H dominant. By L5–L6, they move fluidly between all three.
Hand ≠ Feasibility
No other framework treats Hand as active craft mastery. Whatever second discipline you bring into design, H3 treats it as a multiplier. Hand is making.
Team composition lens
“Are we Heart-heavy and Hand-light?” H3 also diagnoses team imbalance, expanding beyond individual development.
Hiring rubric ready
Each H maps to capability matrix columns. Roles can be profiled as “Heart-primary” or “Hand-primary” before writing the job description.
AI-era relevance
AI compresses Hand (execution) and amplifies Head (systems) and Heart (empathy). The H3 framework explains exactly why human judgment grows more valuable.
Three forces. Three modes.
One practitioner.
Every product lives at the intersection of User, Business, and Technology. Every whole human being thinks, feels, and acts. The H3 model proposes these two triads describe the same territory from different positions: the product triad names the forces a design decision must hold; H3 names the modes of practitioner engagement that meet them. A designer who masters all three modes operates as a complete practitioner. The mechanics are culturally and disciplinarily portable (brand, service, content, research, agency, freelance), even where the surface labels on this page (product triad, IC track, cross-functional partnership) reflect contemporary product-design practice.
Coming into this layer, you likely have strong instincts about your craft but no language for them. You know when your work is good and when it isn’t. You just can’t always explain why, or tell someone else how to get there. You may also be developing broadly without a clear sense of where to go deeper.
After this layer: you will have named your dominant H dimension (the mode you naturally lead with) and used that to identify the specific discipline cluster that rewards your investment most right now. More importantly, you will have one subject on your study list this week. Not a category, not an area of interest. A subject, with a definition of what done looks like after three months of deliberate practice. The practice loop gives you the cadence to keep that going without it becoming another intention you forget by Friday.
“The product triad tells you what you must balance. Hand, Head, and Heart tells you who you must be to balance it. That’s the difference between a framework for ideas and a framework for practitioners.”
There is a long lineage of design definitions H3 builds on. Herbert Simon (1969): “Everyone designs who devises courses of action aimed at changing existing situations into preferred ones.” Victor Papanek (1971): “Design is the conscious and intuitive effort to impose meaningful order.” John Heskett (2002): “The human capacity to shape and make our environment in ways that satisfy our needs and give meaning to our lives.” Charles Eames (1969): “A plan for arranging elements in such a way as best to accomplish a particular purpose.”
H3's contribution is not a new definition. It is a decomposition: the three parts of design, namely human need (Heart), form (Hand), and buildability (Head), become the three modes of the practitioner who does the work. The canon describes the act. H3 describes the actor.
Refs · Simon, H.A. (1969/1996). The Sciences of the Artificial. MIT Press. · Papanek, V. (1971). Design for the Real World. · Heskett, J. (2002). Toothpicks & Logos. Oxford. · Eames, C. (1969). Interview, Louvre.
H3 evaluates and develops people. It maps a designer’s growing capacity across three modes of engagement and bridges directly into career levels, team design, and hiring calibration.
| ↑ Upstream — DESIGN STRATEGIST | ↓ Downstream — EXPERIENCE ENGINEER |
|---|---|
| User research methods (interview, observation, diary study) | Component architecture & design systems |
| Research synthesis & insight framing | Motion & interaction design |
| Journey mapping & service design | Front-end fidelity & code-based prototyping |
| Behavioural science & user psychology | Interaction specification & annotation |
| Systems thinking & problem framing | Design system governance & versioning |
| Product strategy & business framing | Accessibility craft & WCAG compliance |
| Jobs-to-be-done & need framing | Design tokens & naming conventions |
| Emotional design — Norman’s three levels | Developer handoff precision & naming |
| Data literacy & metrics fluency | Prototyping tools at code fidelity (Framer, etc.) |
| DesignOps & process design | UX engineering & front-end fluency |
Problem Framing · Value Shaping
Experience Realisation.
H3 defines what a designer must be. This section names what a designer must do: the three-phase transformation that happens every time a designer engages with a problem within the scope of product usability and desirability.
A designer is a transformer: someone who takes an unstable problem, shapes it through the lens of human value, and brings it into reality as a crafted experience. These three phases are simultaneous responsibilities, and the professional skill is knowing which phase the moment demands.
The phases above and the three layers (Model, Structure, Conduct) meet on every design decision. The cells are alive at every level. What changes across the row is the scope the phase acts on, the time-horizon of consequence, and the mechanism through which the phase happens.
| Model IC making scope |
Structure specialist or lead scope |
Conduct leader scope |
|
|---|---|---|---|
| Problem Framing |
Reading the brief honestly. Spotting when the task description does not match the underlying need. | Reframing or refusing the feature brief. Choosing which of three plausible problems is the real one. | Choosing which problems the team works on. Setting the problem space the org enters next. |
| Value Shaping |
Choosing what matters in a screen. Trading off two equally buildable options on user truth. | Owning the value definition for a feature or programme. Deciding what good means in this product area. | Holding the team’s value standard across products. Setting principles others adjudicate by. |
| Experience Realisation |
Producing the artefact at the standard. Calibrating the craft decision. Shipping the actual thing. | Setting the craft standard others realise against. Building the system that makes others’ realisation faster. | Keeping personal Model alive. Reading shipped work directly. Intervening when the team cannot see what is missing. |
Read down a column to see what your layer asks of you across all three phases. Read across a row to see what the same phase looks like at the level above you. The career legibility H3 has been promising lives in the second reading.
Theoretical grounding (Rittel, Schön, Norman), the L1→L2 cell-shifts (Problem Framing tolerance, Value Shaping coherence, Experience Realisation beginning to teach), and the L2→L3 transition into Structure are developed in Chapter 2 of the book. The matrix is operationalised as Tool 07.
The H3 mapping
to the Product Triad
Each H is both an inner mode of the practitioner and an outer orientation toward a force in the product triad.
Refs · Young, I. (2015). Practical Empathy. Rosenfeld Media. · Young, I. (2022). Time to Listen. · Riess, H. (2018). The Empathy Effect. · Zaki, J. (2019). The War for Kindness. Crown.
Where the H’s meet
The most powerful design decisions live at the overlaps. Each intersection is a quality of practice you develop. A designer who operates only in one H produces work that is technically correct but incomplete. The intersections are where completeness lives.
Worked examples and the diagnostic for which intersection your practice gravitates toward are developed in Chapter 6 of the book.
H3 among existing frameworks
| Framework | Author | Focus | Where H3 differs |
|---|---|---|---|
| Desirability / Feasibility / Viability | IDEO / Tim Brown | Innovation evaluation | Evaluates ideas, not practitioners. No career dimension. |
| Why / How / What (Golden Circle) | Simon Sinek | Leadership & purpose | Org-level; no craft specificity or product triad grounding. |
| Cognitive / Affective / Psychomotor | Bloom’s Taxonomy | Educational learning | Pedagogical framing; not mapped to design practice. |
| Firmitas / Utilitas / Venustas | Vitruvius (~25 BCE) | Architecture triad | Closest ancient parallel. Evaluates artefacts, not people. |
| Design Thinking (5-Stage) | Stanford d.school | Innovation process | Process model; no practitioner development or career levels. |
| Double Diamond | UK Design Council | Design process | Process, not people. No practitioner modes or levels. |
| H3 — Hand, Head, Heart as Design Operating System | Fares Farhan, 2025 | Practitioner development | Maps to product triad, career levels, team composition and growth. |
Beyond T-shaped.
Three dimensions, one person.
The T-shaped designer (broad generalist + one deep skill) was the model of the 2010s. The π-shaped designer adds a critical third dimension (leadership leverage), making it the right profile for the AI era.
Coming into this layer, you likely have a sense of where your career is going but no deliberate architecture behind it. Decisions have been made (which skills to develop, whether to move into management, which projects to take), but mostly in response to what appeared rather than what you chose. You may also be at a crossroads and not yet know it.
After this layer: you will have named your current level honestly, identified your craft-depth direction (upstream or downstream) and your leadership-depth track (people leadership or thought leadership), and written a specific three-month target for your weaker vertical. If you are at Lead level, you will also have run the branch decision card with your manager (not alone), and named your direction. Not deferred it, not kept it open. Named it. That is the output this layer is designed to produce.
“AI makes the I-shaped designer fragile (craft can be replicated) and makes the π-shaped designer formidable (judgment, context, and multiplier effect cannot be replicated).”
On prior art · T-shaped is credited to David Guest (The Independent, 1991) and popularised in design discourse by Tim Brown and IDEO in the mid-2000s (Change by Design, 2009). π-shaped (pi-shaped) has circulated since c.2008 in Accenture and Deloitte human-capital writing, in Linda Liukas’s work on hybrid technical/creative practice, and in Jared Spool’s UX writing; adjacent variants include M-shaped and comb-shaped (multi-deep). H3's contribution is not the π geometry itself but its specific application to designer career architecture: the horizontal bar as design-generalist literacy, craft depth as upstream or downstream mastery, leadership depth as the multiplier axis (people or thought leadership). The shape is inherited; the axes are H3's.
| ↑ Upstream — Design Strategist | ↓ Downstream — Experience Engineer |
|---|---|
| ↑ §01 User research methods & synthesis at depth Execute → Teach: can you write a research brief and lead synthesis with a team? | ↑ §01 Component architecture & design system governance Execute → Teach: can you set the naming convention and review others' tokens? |
| ↑ §01 Behavioural science & psychology applied to design Execute → Teach: can you write a behavioural framing for a design decision? | ↑ §01 Motion & interaction specification at build fidelity Execute → Teach: can you spec transitions that engineers implement without questions? |
| ↑ §01 Quantitative data analysis & metric framing Execute → Teach: can you frame a product problem using metric evidence alone? | ↑ §01 UX engineering & code-level prototyping Execute → Teach: can you prototype a complex interaction in code without a template? |
| ↑ §01 Strategic problem framing & hypothesis formation Execute → Teach: can you write a testable hypothesis that shapes a product roadmap? | ↑ §01 Accessibility engineering & automated testing Execute → Teach: can you write an accessibility audit brief and interpret WCAG violations? |
| ★ Advanced Upstream stakeholder facilitation & alignment Requires the four above as foundation: facilitation without depth is performance. | ↑ §01 Developer handoff precision & front-end fluency Execute → Govern: can you set the handoff standard for the whole team? |
| ♥ People Leadership — Management track | ◈ Thought Leadership — Senior IC track |
|---|---|
| Start here → Feedback delivery Practise SBI framing (Situation–Behaviour–Impact) on real work this week. The bar: someone you gave feedback to changes their behaviour. If you cannot name that person, feedback is not yet landing. |
Start here → Writing a point of view Write 200 words on one design problem you have observed repeatedly. Not a post, a private articulation first. The bar: someone else reads it and immediately understands the problem without asking a follow-up question. |
| 1:1 structure & active listening Build a repeatable 1:1 format (agenda shared in advance, space for their topics first). The bar: your direct reports feel prepared. |
Mentoring through craft Take one mentee. Meet monthly. Track one specific development target together. The bar: their work is visibly different three months later and they can name why. |
| Career development & IDP facilitation Write one IDP with a designer, not for them. The bar: they can describe their own development target without you prompting them. |
Internal speaking & facilitation Present a design decision and its reasoning to a cross-functional audience. The bar: non-designers can explain your reasoning without you in the room. |
| Team health & retrospectives Run one retro where the team generates the actions. The bar: you did not write the action items yourself. They did. |
Conference speaking & community Submit one talk proposal or write one public post. The bar: someone you have never met responds with a question or a counterargument. |
| Hiring process design & calibration Design a calibration rubric for one role before writing the job description. The bar: two interviewers score the same candidate within one point of each other. |
Framework development & IP creation Formalise one repeating pattern in your practice into a reusable model. The bar: a colleague uses it independently without being prompted by you. |
| Org dynamics & stakeholder influence Map the three stakeholders whose approval design needs before work begins. The bar: they are coming to you proactively, not only being informed after decisions are made. |
Cross-industry pattern recognition Study one adjacent field (engineering ops, coaching, product strategy) and write down three things it does that design does not. The bar: one of those three things changes how you work. |
The synthesised
career matrix
The H3 career architecture across six levels: Associate → Senior → Lead → Staff → Principal → Advisor. Each level represents a fundamentally different relationship: with the work, the discipline, and the organisation.
On prior art · The six-level shape with an IC/management fork is the canonical architecture of contemporary tech career literature: Camille Fournier (The Manager’s Path, 2017), Will Larson (An Elegant Puzzle 2019; Staff Engineer 2021), Tanya Reilly (The Staff Engineer’s Path, 2022), and public design ladders at Dropbox, Medium, Figma, Shopify, GitLab. The L1–L7+ numeric convention originates in Amazon/Google/Meta and is now industry-generic. H3 inherits this architecture. The contribution here is the verb set (Follow · Assist · Apply · Enable · Ensure · Direct & Advise), which names each level as an expanding relationship (to the work, to peers, to the discipline, to the field), and the cross-cut of every level against the H-profile and the π-shape.
| L | Verb | IC Title | Management Track | Theme · Scope | Upstream path | Downstream path | AI exposure |
|---|---|---|---|---|---|---|---|
| L1 | Follow | Associate | — pre-branch — | Craft, professionalism. Portfolio built; little shipped work. Function-level problems. | Building research habits | Building prototyping habits | Augments craft tasks |
| L2 | Assist | Senior | — pre-branch — | Communication, craft. Shipped across multiple product areas. Reliable and consistent delivery. | Conducting research; synthesising insights | High-fidelity prototyping; component thinking | Accelerates iteration |
| L3 ★ | Apply | Lead | → Manager Branch decision point. Pivot to people leadership. |
Depth, cross-functional leadership. Owns a discipline. Branch here. | Behavioural analysis; hypothesis framing | Front-end exploration; design system contributions | Co-creates solutions |
| L4 | Enable | Staff | Sr Manager Manages 4–5 reports. Program-level thinking. |
Programs not projects. Deep in 2 skills. Design Strategist or Experience Engineer emerges. | Strategist: data fluency, psychology, upstream influence | Engineer: UX engineering, code-level craft | Directs AI at system level |
| L5 | Ensure | Principal | VP of Design Manages managers. Org-wide leadership. |
Breadth, thought leadership. New patterns and practices. Shapes org process. | Defines strategy methodology; shapes org’s upstream process | Defines design engineering standards; influences product architecture | Governs AI in team |
| L6 | Direct & Advise | Advisor | SVP / CDO Recruiting culture. Company-level accountability. |
Industry-defining. Identifies new value. Attracts talent. Promotes externally. | Shapes industry discourse on strategic design | Shapes industry discourse on design engineering | Sets AI ethics + policy |
Design Individual Contributor
| Category | Tool (late 2025) | What it does |
|---|---|---|
| LLM assistant | Claude, ChatGPT | Thinking partner: pressure-test rationale, draft briefs, generate directions to critique |
| AI research tool | NotebookLM | Upload papers, articles, or research transcripts and interrogate them to extract patterns |
| Wireframe generation | Uizard, Figma AI | Generate wireframe directions from a brief; AI-assisted layout suggestions |
| Ideation exploration | V0, Bolt | Generate UI directions from a prompt, useful as a brainstorming scaffold. Cookie-cutter by default; human judgment determines what to keep, discard, or reframe |
| Craft extension | Claude Code, Cursor | Coding assistants that significantly reduce time from design decision to working proof of concept. Particularly valuable for Experience Engineers building at code fidelity |
| Accessibility audit | Stark (Figma plugin) | AI-flagged contrast, readability, and WCAG issues before handoff |
| User testing | Maze AI | Automated usability testing with AI-synthesised session summaries |
Design Leader
| Category | Tool (late 2025) | What it does |
|---|---|---|
| LLM assistant | Claude, ChatGPT | Draft feedback, OKRs, vision documents, meeting prep, and position statements |
| AI research tool | NotebookLM | Synthesise industry reports, team retrospectives, and competitive research into interrogable knowledge |
| Meeting synthesis | Otter.ai, Fireflies | AI transcription and action-item extraction from 1:1s and team sessions |
| Presentation | Gamma | AI-assisted deck structure and slide generation from a brief |
| Strategy synthesis | Perplexity | Research synthesis for staying across industry trends and competitive landscape |
At L3 the single track ends. From L3 onward, IC and management run as parallel architectures: same level, comparable seniority, different work. The full level-pair architecture is in §08 — Beyond the Branch; the diagnostic for which track fits is at Tool 03 in the back of this page.
The Branch is not a promotion. It is a question about where energy creates more value: through your own output, or through the output of others. Some practitioners reach it earlier, some later, some never cleanly — the Branch is a recognisable pattern, less a principle.
This framework uses Associate rather than Junior. Junior implies a ceiling. Associate implies a beginning.
Two tracks,
four level pairs.
L1 (IC1 Associate) and L2 (IC2 Senior) are pre-Branch — every designer travels the same single track through these levels, building craft reflexes (IC1) and the horizontal bar of breadth (IC2). At L3 the career forks. From L3 to L6, IC and management run as parallel architectures: same level, comparable seniority, different work. The cells below name what each pair looks like in practice.
How to read this: the pairs sit at the same level through different jobs. IC3 Lead and Mgmt1 Manager are both at L3 with comparable scope and seniority, but they hold it through different mechanisms: the IC through discipline and craft ownership, the manager through people and collective output. Neither is above the other. The diagnostic that helps the practitioner choose is at Tool 03 in the back of this page, and in full in Chapter 9 of the book.
Depth, cross-functional leadership. Owns a discipline. Begins craft depth: deepening upstream (research, strategy) or downstream (motion, systems, code-fidelity craft).
3–6 direct reports. Owns the team’s output collectively rather than shipping directly. Begins leadership depth: people leadership and team-scale craft.
“The Branch is not a promotion question. The two tracks are level-equivalent at every step, and choosing the wrong one is costly. The architecture above shows what each path looks like across L3 to L6. The diagnostic that helps the choice is at Tool 03 in the back of this page. The full diagnostic, with counter-tests, the two-year rule for reversals, and a worked example, is in Chapter 9 of the book.”
Upstream or downstream.
Both are needed.
As AI absorbs the middle of the design spectrum, the most defensible positions move toward the two ends. These are specialisation profiles within the existing career architecture.
This matters because of what AI is doing to the shape of design work. The generalist middle (the designer who conducts some research, produces some wireframes, does some visual design, and writes some specs) was the standard of the 2010s. That competency set is increasingly replicable by AI working with a motivated practitioner. What AI cannot replicate is depth at the ends of the spectrum: the researcher-strategist who can frame the problem the organisation didn’t know it had, and the engineer-designer who can build the thing that couldn’t be described until someone made it.
These are not new roles invented for the AI era. They are the natural destination of the π model’s craft depth: the depth that separates a designer who practises a discipline from one who leads from it. The two profiles below describe what that leadership looks like at L4 and above, on each end of the spectrum.
π depth: Behavioural science, data literacy, user psychology, futures thinking
Level range: L4L5L6
- Behavioural science & user psychology
- Quantitative research + data synthesis
- Service design & systems mapping
- Futures thinking & scenario planning
- Business model literacy
- Problem framing & hypothesis writing
Synthesising large qualitative datasets into theme clusters; running behavioural pattern analysis without data science support; stress-testing strategic hypotheses against market data.
The judgment to know which insight matters. The contextual intelligence to frame a problem the organisation can actually act on. The trust earned to influence before anything is built.
π depth: Code-level prototyping, component engineering, design systems authorship
Level range: L4L5L6
- Component architecture & design system authorship
- UX engineering & code-level prototyping
- Motion & interaction specification at build fidelity
- Accessibility engineering & WCAG compliance
- Developer handoff precision & front-end fluency
- Design token governance & naming systems
Generating component variants and states at scale; writing interaction and animation code from a specification; running automated accessibility audits and producing WCAG compliance reports; producing front-end code from design tokens that engineers can use directly.
The taste to judge when a generated component is technically correct but experientially wrong. The systems judgment to set token naming, component APIs, and consistency rules AI cannot arbitrate. The craft instinct to know when the output needs to feel different from what the tool produced.
If the Design Strategist profile made you feel seen, your craft depth is upstream. If the Experience Engineer profile made you feel seen, your craft depth is downstream. If neither felt quite like you, you are still building horizontal breadth, and that is exactly the right place to be at L1–L3. The profiles are not aspirations for every designer; they are destinations for designers who have chosen to go deep.
- —You are more energised by the question before the brief than the work inside it
- —You reach for research and data when others reach for wireframes
- —You want to redirect what the team is trying to solve
- —You are most alive when making, when the work is in your hands and moving
- —You care deeply about how things are built: the gap between design and implementation bothers you
- —You find yourself thinking about systems, tokens, and consistency rules before being asked
Neither profile is complete without leadership depth, whether through people or through thought. The Strategist who cannot lead cross-functional alignment becomes a researcher without influence. The Engineer who cannot mentor the next generation of craft becomes a contractor without leverage. Craft depth is the direction. Leadership depth (people leadership or thought leadership) is what makes it matter.
How AI reshapes
the practitioner.
AI restructures which competencies are defensible, compresses which parts of the skill-building journey, and amplifies which kinds of practitioners become exponentially more valuable. The effect is not uniform: it maps directly to the three H dimensions.
§05 stated the core claim: AI compresses Hand (execution) and amplifies Head (systems) and Heart (empathy). This section is the elaboration: what that means in practice, at each level of the career, for each mode of the practitioner.
-
L1 → L2 transitionCompresses. A motivated L1 with AI reaches L2 output quality faster. The baseline expectation rises: what once took two years may take one. Entry-level craft is not devalued; it is repriced.
-
L2–L3 plateauMost exposed. Established craft, not yet strategic. The generalist middle compresses. The antidote: move toward upstream or downstream craft-depth specialisation faster than you planned.
-
“Follows convention”Partially replaced. Knowing and applying standard patterns is now a baseline AI capability. Necessary as a starting point; no longer sufficient above L2.
-
Problem FramingAmplified. §01.5 (the skill of interrogating a brief before solving it) becomes exponentially valuable. AI produces solutions. Only the strategist decides which problem deserves one.
-
Systems ThinkerAmplified. AI can generate a component. It cannot decide the token naming, component API, or cross-product consistency rules. The systems thinker governs what AI produces.
-
AI Direction (L4+)Created. A new craft dimension: prompting, directing, and governing AI as a design collaborator. Not “using AI tools”; building the design systems AI can build within.
-
BoldnessAmplified. Knowing when to reject AI output and insist on the non-obvious choice. As AI convergence increases, the decision to diverge becomes a professional competency in itself.
-
Empathy at depthIrreplaceable. The insight that comes from being present with users (not from synthesis of prior sessions) cannot be replicated. Value Shaping (§01.5) operates at this level.
-
Fast Learner (redefined)Redefined. No longer “learns new tools quickly.” Now: maintains calibrated judgment as the tools change monthly. Knowing when to trust AI and when not to is core professional practice.
-
Advocacy under pressureIrreplaceable. The capacity to champion a user’s need against a business-only metric, in a room where no one else is, requires a kind of conviction that no model produces.
“AI makes the I-shaped designer fragile and the π-shaped designer formidable, because judgment, context, and the multiplier effect cannot be replicated.”
Systems Thinker
AI can generate a component. It cannot decide the component API, token naming, or cross-product consistency rules. The Systems Thinker governs what AI produces.
AI Direction (L4+)
A new craft dimension: prompting, directing, and governing AI as a design collaborator. Not “using AI tools”; building design systems that AI can build within.
Fast Learner
Becomes: maintaining judgment as the tools change monthly. Continuous calibration of when to trust AI and when not to becomes core professional practice.
L2–L3 plateau
Established craft, not yet strategic: the most AI-exposed position. The antidote: reach the Upstream or Downstream specialisation faster. The generalist middle compresses.
Five domains.
Vision · Direction · Function · Organisation · Cooperation
Design leadership is five simultaneous responsibilities, each operating at a different altitude and requiring a different H at its centre. In my experience, most leaders are effective in two of the five and unconsciously neglect the other three, a pattern observation (offered without claim of measurement). The gaps do not announce themselves loudly; they accumulate quietly until the team loses coherence, the quality that distinguishes a design organisation from a group of designers doing parallel work. The five are the formal surface; a political shadow thread runs alongside them: the informal peer trust, reputation maintenance, and cross-functional credibility that leadership at L4+ spends real time on. This page does not list it as a sixth domain, but the reader operating at that scope should not be surprised to feel it.
Coming into this layer, you are likely leading by instinct across five responsibilities you may not have names for. Some things are working. Some have gaps you can feel but haven’t been able to locate precisely enough to act on. You may also be spending energy on the domains you are already strong in while the limiting domain sits quietly in the background.
After this layer: you will have rated your own practice across all five management domains (Vision, Direction, Function, Organisation, Cooperation) and identified the one domain that is currently limiting your team’s coherence. You will have run the Quarterly Leadership Cycle (Tool 04) end to end: Movement 1 assesses the five domains, Movement 2 uses the Focus Selection Protocol to confirm the choice, and Movement 3 records the Focus Card: three tasks, one owner each, one progress signal each, with a date to share it with your team. That card is not a planning document. It is a commitment made visible. That is the output.
The five domains give that accumulation a structure you can see, assess, and act on. When all five domains are in a healthy state, the team has direction, craft standards, a functioning operation, people who are growing, and a seat at the business table. That combination is what coherence feels like from the inside. The goal of this layer is to help a design leader build and sustain all five domains, including the ones outside their natural pull.
The five domains follow a dependency chain. Vision anchors everything: without strategic direction, all other domains are improvisation. Direction is the creative expression of that vision: the principles and quality bar that flow from knowing where the practice is going. Function designs how the practice operates: the structure, process, and craft standards that flow directly from the direction you have established. You design the operating model before you hire to staff it. Organisation realises the function with the right people: hiring for the ensemble’s gaps, developing talent, and shaping culture around a process that already exists. Cooperation earns external influence last: a design practice with clear vision, established direction, a functioning operation, and the right people has something credible to bring to cross-functional partners. Without those foundations, cooperation is advocacy without weight.
Direction, Goals & Roadmap
Map the current state · Set practice direction & goals · Build the design roadmap · Govern design programmes · Track practice health.
Principles, Standards & Craft
Establish design principles · Build the visual & interaction language · Champion the user · Shape the end-to-end journey · Give decisive creative guidance.
Structure, Tools, Process & Environment
Design the team structure · Establish the design process · Uphold craft standards · Create the conditions for good work · Build a learning culture.
People, Culture & Growth
Hire for the ensemble’s gaps · Shape the team’s working culture · Give honest and timely feedback · Grow each designer’s potential · Develop the next layer of leadership.
Business, Partners & Influence
Connect design to business direction · Build trust with key partners · Champion design-led problem solving · Bring design to org-wide problems · Build design’s external presence.
How to assess domain health
★★★ Effective — functioning; maintain.
★★ Meeting Some Expectation — gaps; plan this quarter.
★ Needs Improvement — blocking; act now.
- Each domain draws on a distinct body of knowledge. Vision requires strategic thinking and OKR design. Direction requires design criticism, taste development, and emotional design theory. Function requires process theory and systems thinking applied to operations. Organisation requires adult development frameworks, feedback models, and coaching approaches. Cooperation requires stakeholder influence theory and cross-functional leadership dynamics. Your lowest-rated domain in the assessment is your study area.
- One book per domain you are actively developing: not a reading list, a discipline. Vision: Good Strategy Bad Strategy (Rumelt) or Measure What Matters (Doerr). Direction: The Shape of Design (Chimero) or Emotional Design (Norman). Function: The Goal (Goldratt) or Team Topologies (Skelton & Pais). Organisation: The Manager’s Path (Fournier) or Radical Candor (Scott). Cooperation: Influence Without Authority (Cohen & Bradford) or The Trusted Advisor (Maister).
- Each domain has an adjacent field that has already solved problems design is still improvising around. For Vision, follow strategy consultants and design maturity researchers. For Direction, read design criticism: reviews of products and artefacts written by practitioners with strong opinions; it trains the vocabulary of taste. For Function, study how engineering and product operations are run; adjacent disciplines have solved coordination problems design treats as unique. For Organisation, look at coaching and adult learning theory; the ICF coaching competency framework is a useful lens even outside formal coaching. For Cooperation, study how the best product leaders build cross-functional influence; the mechanics of trust across disciplines are the same regardless of who is building it.
- Score all five domains. Identify your Needed domain, the one blocking everything else.
- Name one specific leadership task within that domain that is currently absent or broken. Use the reference below to identify it.
- Write the single next action that would move it from Needed to Improve: owner, deadline, definition of done.
- Is your Needed domain still Needed from last quarter? If yes, what specifically stopped you?
- Can one of your designers name the domain you are currently focused on without being told? If not, your focus is not visible enough to create organisational movement.
| Domain | H3 Axis | Key leadership tasks | Signal of Needed state |
|---|---|---|---|
| Function | Head + Hand | Design workflow · Quality assurance · Tooling & environment · Knowledge exchange | Work is inconsistent; no shared process or quality bar |
| Direction | Heart | Design principles · User perspective advocacy · End-to-end experience · Innovation drive | Team has no shared view of what good looks like |
| Vision | Head | Situation assessment · OKR & goal setting · Strategy & roadmap · Measuring performance | No roadmap; design work is reactive and ad hoc |
| Cooperation | Head + Heart | Business alignment · Stakeholder alliances · Design reputation · Culture promotion | Design is invisible in planning and strategy conversations |
| Organisation | Heart | Team culture · Talent growth & IDP · Recruiting & onboarding · Feedback & guidance | Attrition risk; designers feel stuck or unseen |
- —Which areas and aspects are doing well?
- —Which aspects are working not so well?
- —What are the biggest pain points?
- —Do we have blind spots?
The management matrix needs
an assessment mechanism.
Each of the five management domains contains five specific leadership responsibilities. The three-star rating system gives the leader a simple, honest mechanism to assess where they currently stand, and where to focus next.
The assessment is not a performance review. It is a current state map: run quarterly, shared with the team, used to feed the Focus Selection Protocol that follows. The honesty of the rating is more important than the comfort of the score.
Rate each aspect honestly. A two-star rating is not failure. It is information. The goal of the assessment is not to produce a perfect score but to produce an honest map that makes the next decision obvious.
4 Questions · 15 Minutes · One Domain
Once the assessment is complete, this protocol converts it into a single quarterly focus. A comprehensive framework without a selection mechanism produces paralysis; this is the mechanism.
“The assessment tells you where you are. The protocol tells you what to do next. Neither works without the other.”
Culture and capability
need a visible link.
Culture values and capability dimensions describe the same person from two different angles. Without a visible link between them, culture values read as aspirational HR language rather than functional performance predictors, and hiring decisions lose their grounding in what the role actually requires.
Root cause: Culture values capture who you are. Capability dimensions capture what you can do. Without the bridge, the link between them is invisible, making it impossible to explain why culture values have practical consequences for career development and hiring.
| Culture Value | Mechanism | Capability Dimensions Activated | H3 Axis |
|---|---|---|---|
| Baseline — universal qualities every designer in the team needs | |||
| Humble | Receives critique without defensiveness; enables mentoring | Follows ConventionMentoring | Heart |
| Passionate about Design | Maintains standards when no one is checking | Strives For QualityStrong AestheticsBoldness | Hand |
| Collaborative | Builds trust that enables design to influence earlier | FacilitationMentoringBusiness Sense | Heart + Head |
| Clearly Communicates Ideas | Enables design thinking to travel beyond the design team | FacilitationProduct SenseSimplify Complexity | Head + Heart |
| Curious | Drives self-directed learning beyond the brief | Fast LearnerUX & InteractionEngineering Skills | Hand |
| Added Value — culture-specific qualities that deepen the practice | |||
| Integrity | Maintains quality standards under deadline pressure | Strives For QualityFollows ConventionSystems Thinking | Head |
| Respect | Creates psychological safety in critique; enables risk-taking | MentoringFacilitation | Heart |
| Positive Impact | Keeps user benefit as north star vs. business-only metrics | UX & InteractionBusiness SenseProduct Sense | Heart + Head |
| Growth | Actively seeks feedback and applies it | Fast LearnerDeliveryPrototyping | Hand |
| Operational Excellence | Applies craft rigour to process, not just output | DeliveryFollows ConventionSystems Thinker | Head + Hand |
Capabilities with no culture value driver → must be hired for
Strong Aesthetics
Visual taste is not reliably cultivated by culture values. Screen for it in portfolio, not values interview.
Systems Thinking
Learnable but requires deliberate practice beyond day-to-day project work. A structured programme covering information architecture, component-system governance, service blueprinting, and feedback-loop analysis is needed. Anchor readings: Meadows, Thinking in Systems; Senge, The Fifth Discipline.
Engineering Skills
Curiosity helps; but the base level must be screened at hire for roles where it is required. Curiosity alone won’t close a large gap fast enough.
Design needs a delivery process
that matches its ambition.
The H3 Function domain calls for a defined design workflow, structured work streams, collaborative design practice, and quality assurance. H3 Operations is one worked configuration: a six-phase Product Design & Development workflow — Discover · Define · Design · Deliver (the first four D’s adapted from the Design Council’s Double Diamond, with Design in place of their Develop) extended through Develop and Deploy, aligned with Scrum cadence and a phase-by-phase RACI accountability matrix.
Why this exists: H3 defines who the designer is and how they lead. It does not prescribe the delivery process that runs beneath leadership. Without a structured workflow, the Function domain remains aspirational. The chapter offers one shape of that workflow; agency, freelance, solo in-house, and non-product practices substitute the Accountable role and adjust the cadence. The RACI logic travels; the labels do not.
Six Phases at a Glance
Problem Space (1–2) → Solution Space (3–4) → Concrete Space (5–6).
The full configuration — RACI legend, Scrum goal hierarchy, per-phase activities and deliverables, and the framework’s mapping to the five H3 Management Domains — is summarised in Appendix D of the book and published in full as the H3 Operations sister document.
Seven instruments.
One framework in practice.
The tools below make the H3 framework operational. Each translates a layer or section into a concrete act. Run them in sequence, in isolation, or return as the situation demands.
Rate your current strength across fifteen capability areas, five per H. The output: your primary H, your development priority, the gap, and your dominant intersection (Purpose, Presence, Precision, or Mastery). Calibration note: self-rating tends to inflate on the leadership side (Head, Heart); pair this with a peer or manager score for a truer read.
| Capability area | Rating 1 = not yet · 5 = consistently strong |
|---|---|
| ✦ Hand dimension | |
| Craft quality — my output is consistently at a high standard | |
| Technical depth — I understand how what I design gets built | |
| Systems thinking at craft level — I see patterns, reuse, and naming opportunities | |
| Prototyping fidelity — I can prototype at the resolution the problem needs | |
| Taste and judgment — I make and defend considered craft decisions | |
| ◈ Head dimension | |
| Problem framing — I question the brief before I begin solving | |
| Strategic thinking — I connect design decisions to business outcomes | |
| Systems thinking at org level — I see how decisions compound over time | |
| Research synthesis — I turn observations into insight and direction | |
| Data fluency — I read and use quantitative data in design decisions | |
| ♥ Heart dimension | |
| User empathy — I design from specific human need | |
| Facilitation — I can lead a room to a shared understanding | |
| Mentoring — I develop others' capacity alongside my own output | |
| Stakeholder trust — I have built relationships that give design influence early | |
| Team sensing — I notice what the team needs before they say it | |
Map your shape across the three bars of the π model: the horizontal bar of breadth, craft depth, leadership depth. Add level (current vs target) and the one capability you will move this quarter. The shape that emerges is more useful than any single score in any single row.
Run this with your manager, not alone. The conversation is half the tool. Score each signal on a three-point scale — W (weak), P (present), S (strong) — based on your behaviour over the past twelve months. Observed evidence, not aspiration. The shape across all six matters more than any single one.
Each signal carries a counter-test that distinguishes the real signal from its closest look-alike. Read the counter-test before scoring — that is where most self-assessments quietly inflate.
Three movements, one cycle. Assess the five domains, select one focus, record the commitment. Run this every quarter; the assessment and the card lock together. The AI-era overlay (Compress · Amplify · Create) runs in parallel as a personal capability plan (see the T04 worksheet, page 6).
Three-star rating across the five management domains. ★ = needs improvement · ★★ = Meeting some expectation · ★★★ = effective. The honesty of the rating matters more than the comfort of the score.
| Management domain | Rating |
|---|---|
| Vision : strategic direction and intent | ★★☆Meeting some expectation |
| Direction : craft standards and creative judgement | ★★☆Meeting some expectation |
| Function : process, tooling, delivery discipline | ★☆☆Needs improvement |
| Organisation : team composition, levels, hiring | ★★☆Meeting some expectation |
| Cooperation : cross-functional alignment and stakeholder trust | ★★☆Meeting some expectation |
Full aspect-level assessment (five responsibilities per domain) lives in §12 Domain Assessment. Use that if a single domain needs deeper resolution before you select.
Four questions, fifteen minutes. The protocol converts the assessment into one domain.
The Focus Card. Posted where the team can see it before the quarter begins, reviewed in the first week of the next quarter.
Rate your team’s collective capacity: what the team can actually produce together, not the sum of individual ratings. A team of five balanced designers still has a distinctive H3 shape; the audit surfaces it before the next role is written.
List each designer and mark their dominant H. Dominant = the H they reach for first when the work is ambiguous.
Three team-scale capacities. Score against the team’s last two quarters of actual work, not against the sum of individual H3 ratings. Rubric: 1★ scarce (only the leader holds it) · 2★ present but fragile (one-to-two people, overcommitted) · 3★ well-distributed (multiple practitioners hold it reliably).
The audit precedes the role profile. You cannot specify what the next hire should fill until you have read the current distribution honestly. Output feeds directly into role-profiling with Tool 02 applied at the leader’s scale.
Each culture value activates capability dimensions through a specific behavioural pattern — the mechanism. Without the mechanism named, the value is a poster. With it named, the value becomes an instrument the team can hire against, coach against, and review against.
Use in hiring (does this candidate’s value profile predict the capabilities the role needs?) and in performance conversations (is this designer’s value translating into the capability the role requires?). The bridge audits the translation from value to capability. Coach the translation; do not audit the value.
| Value | Mechanism · what the value does under work conditions | Capabilities activated | H3 axis |
|---|---|---|---|
| Baseline · universal qualities every designer needs | |||
| Humble | Receives critique without defensiveness; enables mentoring | Follows Convention; Mentoring | Heart |
| Passionate about Design | Maintains standards when no one is checking | Strives for Quality; Strong Aesthetics; Boldness | Hand |
| Collaborative | Builds trust that enables design to influence earlier | Facilitation; Mentoring; Business Sense | Heart + Head |
| Clearly Communicates Ideas | Enables design thinking to travel beyond the design team | Facilitation; Product Sense; Simplify Complexity | Head + Heart |
| Curious | Drives self-directed learning beyond the brief | Fast Learner; UX Interaction; Engineering Skills | Hand |
| Added-value · culture-specific qualities that deepen the baseline | |||
| Integrity | Maintains quality standards under deadline pressure | Strives for Quality; Follows Convention; Systems Thinking | Head |
| Respect | Creates psychological safety in critique; enables risk-taking | Mentoring; Facilitation | Heart |
| Positive Impact | Keeps user benefit as north star over business-only metrics | UX Interaction; Business Sense; Product Sense | Heart + Head |
| Growth | Actively seeks feedback and applies it | Fast Learner; Delivery; Prototyping | Hand |
| Operational Excellence | Applies craft rigour to process, not just output | Delivery; Follows Convention; Systems Thinking | Head + Hand |
Locate yourself in one cell of the matrix below. Then read up the column above you and across the row at your level. The cell you operate in names the work; the column above names what your seniors hold; the row across names what the same phase looks like at the levels you do not yet sit at. Use: 1:1s, career conversations, hiring debriefs, IDP writing, team retros where coherence has slipped.
| Model | Structure | Conduct | |
|---|---|---|---|
| Problem Framing | Reading the brief. Spotting when the task does not match the need. Asking the clarifying question. | Reframing or refusing the brief. Writing the hypothesis. Choosing which problem is the real one. | Choosing which problems the team works on. Saying no to problems that look strategic but are not. |
| Value Shaping | Choosing what matters in a screen. Local trade-offs on user truth. | Owning the value definition for a feature or programme. Deciding what good means here. | Holding the team’s value standard across products. Setting principles others adjudicate by. |
| Experience Realisation | Producing the artefact at the standard. Calibrating the craft decision. | Setting the craft standard others realise against. Building the system that makes their realisation faster. | Keeping personal Model alive. Reading shipped work directly. Intervening when the team cannot see what is missing. |
- Down your column. Name one cell where your work currently shows the layer’s shape. Name one cell where it does not.
- Across your row. Pick the phase your work most often gets pulled into. Describe what the cell at the layer above you looks like in concrete terms. If you cannot describe it, that is the legibility gap.
- Bottom-right cell. If you are at Conduct scope: when did you last operate in Experience Realisation directly — reading shipped work, intervening in craft? If the answer is “not this quarter,” the calibration leg of leadership is weakening.
- Standing vs reading. A practitioner whose perception runs ahead of their standing is the practitioner the next level is looking for. Note the cells where you can read the work but cannot yet act on it.
The matrix is mapped in full, with the L1→L2 and L2→L3 transitions walked cell by cell, in Chapter 2 of the book.
How the three layers
connect and reinforce each other
The Model informs The Structure
The branch decision at L3 Lead is a H3 question: where does the practitioner’s energy create more value, through their own output (IC track) or through the output of others (management track)? The dominant-H heuristic is too coarse, see Chapter 9 for the six energy signals. The π specialisation profiles (Strategist / Engineer) map directly to upstream-leaning and downstream-leaning H3 orientations.
The Structure informs The Conduct
The Focus Selection Protocol should be run by someone operating at L4 or above, the level where systems thinking (Head) is strong enough to diagnose domain gaps. The Values × Capability bridge is used at hire (entry levels) and promotion (senior levels).
The Conduct reinforces The Model
The five management domains (Vision, Direction, Function, Organisation, Cooperation) each correspond to a dominant H3 axis: Vision = Head; Direction = Heart; Function = Head+Hand; Organisation = Heart; Cooperation = Head+Heart. The management leader who understands H3 knows which H they are applying in which domain.
H3 — Hand, Head, Heart as Design Operating System, authored by Fares Farhan based on 25 years of digital product design practice with software engineering background. Career levels synthesised from the design team structures I encountered across startups, scale-ups, and Fortune 500 companies, observed, not borrowed.
This work is licensed under Creative Commons Attribution 4.0 International. Share, remix, and apply with credit to Fares Farhan and a link to the license.
The longer arc
of an inherited idea.
The Lineage section opens with Pestalozzi as the origin. The triad did not stay there. It travelled across two centuries, taking on different vocabulary in each domain it landed in: pedagogy, craft education, experiential learning, public extension, and contemporary management writing. This appendix collects the trail in one place for the reader who wants to follow it.
John Dewey carried the Hand dimension forward: learning by doing, experience as the primary teacher. Rudolf Steiner made “Head, Heart, Hand” (Denken, Fühlen, Wollen) the explicit institutional motto of Waldorf education from 1919 onward, with the triad taught as the architecture of a whole human being. The Bauhaus, opening the same year in Weimar, fused craft (Hand) with intellectual rigour (Head) and social purpose (Heart) under Gropius’s manifesto. Kurt Hahn built the triad into Outward Bound (1941) as the foundation of experiential education. The American 4-H youth movement (Head, Heart, Hands, Health, 1902) carried it into a public-domain extension that reached millions. In contemporary management writing (Daniel Pink, Dave Ulrich, Simon Sinek), the phrasing returns under different labels.
H3 does not invent the triad. It inherits a 250-year-old formulation and applies it to a problem it has not yet been applied to: the whole practitioner in the design profession.
References · Pestalozzi, J.H. (1801). Wie Gertrud ihre Kinder lehrt [How Gertrude Teaches Her Children]. · Brühlmeier, A. (2010). Head, Heart and Hand: Education in the Spirit of Pestalozzi. Cambridge: Sophia Books. · Steiner, R. (1919). Stuttgart Waldorf lectures. · Dewey, J. (1916). Democracy and Education. · Gropius, W. (1919). Bauhaus Manifesto. · Hahn, K. (1941). Outward Bound founding principles. · 4-H (1902). Head, Heart, Hands, Health — USDA Cooperative Extension.