H3 Design OS

Before a designer can model their practice, they need a clear answer to what they are modelling. This chapter challenges the design industry's habit of defining designers by the tools they happen to use, the deliverables they produce, or their position on the org chart, the person between PM and engineer.

Instead, it establishes a first-principles definition: a designer is a person who translates human need into form that can be built. That definition has three components, human need (Heart), form (Hand), and buildability (Head), which seeds the H3 framework: Hand, Head, and Heart as three simultaneous modes of engagement that combine to answer the three questions any product must balance: Can I make it? Can I think through what it should be? Can I feel who it’s for?

That is the move that distinguishes H3 from every other framework in design practice. Other frameworks evaluate ideas. H3 develops the practitioner who produces them. The chapter draws four structural distinctions that make this explicit: what separates H3 from the frameworks that came before, and why those differences matter for the practitioner.

A note on scope: H3 applies across the range of organisational contexts designers actually work in: from design-mature product organisations, to teams where design is new or contested, to the sole designer inside a non-design company. The framework’s mechanics do not require a supportive environment. They describe the practitioner’s own development, which proceeds at a different rhythm depending on how much room the organisation gives.

Chapter 1 defined what a designer is. This chapter is about what a designer does.

Every time a designer engages with a product or experience problem, they move through three phases held in parallel: Problem Framing (Head-dominant), Value Shaping (Heart-dominant), and Experience Realisation (Hand-dominant).

A junior designer defaults to Experience Realisation because making is visible and feels like progress. A senior designer has learned the cost of building the wrong thing beautifully, and the discipline to stay in Problem Framing longer.

This chapter is the work model that sits beneath the H3 identity: who you are (Hand · Head · Heart) is the model. What you do (Frame · Shape · Realise) is the practice.

A note on vocabulary: H3's mechanics are intended to be culturally and disciplinarily portable, but its surface language, cross-functional partnership, IC track, advocating upward, reflects contemporary product-design practice. Readers in different cultural settings, in agencies, or in non-product disciplines (brand, service, content, research) should translate the behaviours into their own working context. The labels are a contemporary product-design register, and the framework lives beneath them.

The Pi (π) model is the framework's evolution beyond the T-shaped designer, with three dimensions: the horizontal bar of breadth (design generalist literacy), the first vertical bar of craft depth (upstream or downstream specialisation), and the second vertical bar of leadership depth (people or thought leadership). The AI era makes the Pi profile a necessity: AI has made the purely horizontal designer fungible.

One exception: the hyper-specialist I-shape, the deep research methodologist, the motion designer who wrote the accessibility guidelines for animation, the specialist whose single depth is itself beyond AI's reach, remains resilient. AI fragility comes from breadth with shallow depth, the thing AI most readily replicates, more than from the shape of the profile itself. Going wide with thin legs is the risk. Going deep in one rare place is not.

The two-page career map most designers do not have. The IC track (Associate › Senior › Lead › Staff › Principal › Advisor) sits next to the Management track (Manager › Sr Manager › VP of Design › SVP / CDO), with the Branch at Lead clearly marked. Each level pairs with the H3 profile that matches it, the scope it covers, and the verb that defines its day. The two tracks are not a hierarchy. They are two different answers to the same question about where your energy creates the most value.

On compensation: level is also the market language for pay, scope, and expectation, with wide variance across regions, organisations, and economic cycles. H3 deliberately does not prescribe numbers, because they date fast, but readers should treat their level as both an internal capability description and the external language a market uses to price it.

On plateaus: most of any career is spent between levels: the twelve-to-twenty-four-month stretch where the work expected of the next level is only partly within reach, while the work expected of your current level no longer fully challenges you. The plateau is where depth compounds into the capability that makes the next level possible. This chapter treats it as part of the work, not an absence of it.

At the Lead level, every designer reaches the same fork: Individual Contributor or Manager. This is the most consequential career decision a designer makes, and the one made with the least systematic support. This chapter provides the Branch Decision framework: six behavioural signals (three per track) that ground the choice in observed patterns rather than aspiration. It also connects the Branch directly to the designer's H3 profile: the IC track deepens Hand and Head, while the management track pivots to Heart and Head.

A note on placement: the Lead anchor reflects the common pattern at large product organisations. In smaller teams, agencies, and solo practices, the Branch may appear earlier, later, or at a level that does not map cleanly to Lead. Use the six signals in this card, not the level label, as your guide.

A third branch exists and should be named: leaving design altogether, into product management, research leadership, strategy, founding, or an adjacent creative discipline. Hand, Head, and Heart are transferable capabilities. Walking out of the designer title with them intact is a legitimate path in its own right.

Every designer I have spoken to in the last two years has asked a version of the same question: what does AI do to my career? The useful answer is less a prediction about tools than a map of where AI compresses the work, where it amplifies it, and where it creates competencies that did not exist three years ago, read against the career architecture Chapter 7 built.

The chapter's central asymmetry: AI compresses the generative middle of Hand, amplifies Head, and makes Heart-primary skills the hardest for AI to compress. Hand loses the production layer where craft reps used to accumulate over years. Head gains value because systems thinking and architectural judgement matter more when the rest of the work is cheap. What holds in Heart, and why, the rest of this chapter develops.

Heart survives on two legs. The first is courage: advocating for a user against business pressure in a room where no one else will. The second is empathy at depth: the insight earned by being present in the interview room, which synthesising prior sessions cannot reach. Neither is pattern recognition, which is precisely the place AI is now strong. The specific moves Heart makes in a design decision are not what AI is generating. Whether AI will eventually reach them is a separate question, left open.

This is written from a senior position. The response to AI compression, from here, is to specialise upward, into work AI cannot yet reach. The harder companion question is how the field stays open to practitioners who have not yet reached the rung this framework starts from. If AI compresses the middle of Hand, it also compresses the path designers used to walk through that middle to reach senior practice. The question is named. It is not resolved.

Design leadership is five distinct responsibilities held simultaneously: Vision, Direction, Function, Organisation, and Cooperation. H3 gives the leader a way to see which H each domain calls for. Vision is Head. Direction is Heart. Function is Head+Hand. Organisation is Heart. Cooperation is Head+Heart.

In my experience, most leaders are strong in two of the five and unconsciously neglect the other three. A pattern observation, not a research finding, offered here as a provocation rather than a statistic. The gaps do not announce themselves loudly. They accumulate quietly until the team loses coherence.

A sixth thread runs alongside these five and is often the most time-consuming: the political work of keeping the design function viable: protecting headcount, surviving re-orgs, managing up through leadership changes, advocating for design's place in the budget. H3 treats this as the shadow cast across all five domains rather than a sixth domain in its own right, but it is real and this chapter names it rather than pretending it is absent.

Leadership is also relational: who you are actively mentoring and who is actively mentoring you are part of the Conduct layer, not adjacent to it. The designers you grow and the voices you let reshape your thinking are leadership work.

Knowing five domains exist does not tell you which one to fix this quarter. The Three-Star health model and the Focus Selection Protocol answer that: a four-question process that converts a domain health assessment into a concrete quarterly action plan. It also addresses the hardest question in design leadership: how do you maintain excellence across all five domains simultaneously when resources, time, and energy are always finite? The answer: you do not. Growth is not simultaneous on all three Hs or all five domains. Most quarters, one leads and the others maintain.

On calibration: self-assessment is a mirror, and mirrors reflect what you want to see. Leaders consistently over-rate Vision and Cooperation and under-rate Organisation. Rate the five domains yourself first, then compare with a trusted peer, your manager, or a cross-functional partner who sees the team from outside. Where your rating and theirs disagree is usually where the work actually sits.

On stalls: the chapter also names three patterns that describe when H3 itself stops producing progress, the plateau under H3 (next-level signals partially present, usually a single-H blind spot), blocked by the room (the practitioner develops but the organisation has no slot or language for the level above), and dimension collapse (the dominant H has become the only H). Naming the stall pattern is the first move. The Focus Selection Protocol is the second.