Automation × Gen AI × Skills — Individual Navigation

The Double Hexagon at a glance. This walks one focal question through Hexagon 1 — Align, Scan, Sense-make — using Three Horizons to read the landscape and Futures Wheel to trace one branch outward. The example is deliberately framed at the individual scale, not the systemic one. We're not modelling the labour market. We're sitting next to one specific person, looking at the same horizon with her, and asking what does she keep, what does she drop, what does she combine? That's a different question than "what will AI do to jobs," and it produces different answers.

How to read this example

Each step is signposted:

─── STEP N of 6 ─── HEXAGON 1 · <PHASE> · <TOOL> ───

You can read straight through or pick one step. Each ends with a Try it yourself prompt — copy-pasteable to an AI chat, or usable as a coaching/peer exercise.

A framing note. This is not a competitiveness exercise. We're not asking how Maya beats AI. We're asking how she reads the changing terrain and notices what's already shifting under her feet — what she stays close to, what she lets go of, what she combines into something the old labels don't quite name. Three Horizons is good at this precisely because it isn't an action plan. It surfaces what's already changing so the navigator doesn't get caught flat-footed.

Confidence note. Statistics about today's freelance market come from referenced studies. The character of Maya is composite — drawn from publicly documented working conditions of mid-career freelance illustrators, not modelled on any one person. The futures branches are speculative traces, not predictions.

Why this topic, why these tools

Most public conversation about "AI and jobs" sits at two scales: the systemic (the labour market, GDP, education policy) or the panicky (will I lose my job?). Both miss the texture that decides what actually happens in someone's life. The first is too abstract to be useful for an individual; the second is too narrow to be useful for thinking.

The individual-navigation frame asks: what does the world look like through this person's eyes, in this domain, on this timeline? That puts the question on the same surface as the decisions she's actually making — what client to take, what skill to invest in, what equipment to buy, who to stay close to professionally.

Three Horizons is the right tool for the scan because it forces you to hold three things in your hand at once: a fading present, an emerging future, and the messy middle that's neither. Futures Wheel then takes one item from that middle and chases consequences outward, so you don't just have categories but also chains of "if-then" you can argue with.

Focal question: How might Maya — a 34-year-old freelance illustrator in Mexico City — read her own next 7 years?

A note on framing. We chose 7 years. Three is too close (it's the calendar); 15 is too far (it's a magazine article about the future). Seven is roughly the length of one career-stage shift: a junior becomes a mid; a mid becomes a senior; a senior either becomes a partner / runs a studio / quits / something else. It's long enough that the question can't be answered by next month's plans and short enough that it's still recognisable as Maya's life.

STEP 1 of 6 · HEXAGON 1 · FRAME · Sketch the navigator

We sketch Maya. The discipline here is concreteness. "An illustrator" can't be navigated. Maya can.

Maya, 34. Lives in Roma Sur, Mexico City. Mid-career freelance illustrator.

Roughly 60% of her income is editorial: monthly retainers with two Mexican magazines, occasional New York and Berlin clients picked up via her agent. 25% is small-press book covers (mostly literary fiction, two LGBTQ+ small presses). 15% is what she calls "the weird stuff" — band posters, friend's wedding portraits, a recurring tarot deck commission.
Trained at La Esmeralda (ENPEG). Has a small print studio in her apartment building, shares with a friend.
Earnings: roughly 380,000 MXN/year in 2025 (≈ US$19k), which puts her comfortably middle-class in Roma Sur but precarious by the standards of her foreign clients.
Uses Procreate and a Wacom Cintiq. Tried Midjourney briefly in 2023, doesn't use it for client work.
Single, no kids. Cat. Mother lives in Querétaro.
Started teaching a workshop on editorial illustration at a private design school last year — 6 hours / week, decent supplement.

The focal question is not "will Maya lose her job to AI?" The focal question is what Maya should attend to over the next 7 years that she isn't already attending to.

This is the work of Align/Frame: name the person, the question, the timeframe, and the unit of decision — and notice that we've quietly committed to attention as the unit of decision, not strategy or competitiveness. That commitment will shape the rest of the example.

Try it yourself

Sketch your navigator — yourself, or a specific person whose decisions
you want to think alongside. Include:
- Age, location, family, household economics
- Current revenue / income shape (3–5 sources, % of each)
- Tools and infrastructure they actually use
- One thing they tried once and didn't stick with
- One adjacent thing they're already doing on the side
Then write the focal question in this form:
"How might <NAME> — <ROLE / DOMAIN / LOCATION> — read their own next
<N> years?"
Avoid framing the question as "how do they survive / compete / win".

STEP 2 of 6 · HEXAGON 1 · SCAN · STEEP+++ through Maya's window

We scan the same world everyone else scans — but through Maya's window, not through a labour-economics one. The signals are the same; the salience is different.

Technological

Image-generating models (Midjourney, Stable Diffusion, Adobe Firefly, DALL-E successors) produce client-acceptable mid-quality illustration at near-zero marginal cost. The 2024 Brookings analysis found a 17% decrease in image-creation freelance job posts after major image-model releases. (1)
Adobe Firefly is integrated into Photoshop/Illustrator — tools Maya may end up using even if she "doesn't use AI."
LLM-driven brief-writing means editorial clients now produce more, finer-grained creative briefs. The relationship between art director and illustrator is being mediated by a third interlocutor.

Economic

The mid-market for illustration commissions ($150–$800 range) has collapsed fastest. Editorial fees on the higher end (named bylines, art-directed series) have held steadier. (2)
Some freelancers in design who adopted AI early have seen 40–60% higher hourly rates, per Upwork data — but the sample skews toward those who survived. (2)
Mexican peso volatility means cross-border editorial work is doubly affected — by rate compression and by FX swings.

Social / Cultural

"Hand-drawn" / "handmade" / "human-made" labels are emerging as commercial differentiators in print and merch — partly counter-positioning against AI imagery.
Editorial publications have begun to publish authorship/AI-use policies. Some forbid AI imagery; some require disclosure; some are silent.
Reader fatigue with AI-aesthetic visual sameness is being reported anecdotally — not yet measured well.

Political / Legal

Copyright cases around training-data and AI-output liability are moving through US, EU, and Chinese courts — outcomes uncertain through at least 2027.
Mexican copyright law evolution lags but tracks. Maya's downstream rights when her work is used to train models are uncertain.

Generational / +++

Younger illustrators (graduating in the 2020s) skew either AI-native or AI-refusing — fewer are in the middle.
Maya's cohort (early thirties, ~10 years out of school) is in the peculiar position of being too experienced to retool fast but not senior enough to coast on relationships. This cohort effect matters.

Gap check. We've scanned a lot in technology and economics; less in cultural meaning of authorship, which is a register Maya would feel acutely. A real scan would lean harder there.

Try it yourself

Scan the world your navigator sits in, through *their* window. List
5 signals under each STEEP+++ category, each tagged with what changes
*for them* if it holds. Resist the impulse to scan the labour market
in general — scan the part of the world that meets their week.

STEP 3 of 6 · HEXAGON 1 · SENSE-MAKE · Three Horizons

Read the landscape: what's fading (H1), what's emerging (H3), and the contested middle (H2).

Now the central tool: Three Horizons. We draw three rough S-curves on one sheet — H1 (today's dominant pattern, fading), H3 (the emerging system, growing), H2 (the messy in-between). Then we place signals on whichever curve they belong to and notice where we disagree with ourselves.

The mindsets associated with each horizon — managerial for H1, entrepreneurial for H2, visionary for H3 — are all needed. Three Horizons is not a "transition to H3" exhortation. It's a way of holding the three at once and asking what each demands. (3)

H1 — The pattern that's been working (and is fading)

What Maya's current livelihood is built on:

Editorial illustration commissioned by named art directors at named publications, with byline credit and rights specified.
Trust networks: the same 8–10 art directors across 5 publications and 3 small presses.
A craft signature recognisable enough that someone choosing illustrators says "I want a Maya cover."
Slow client cycles: 2–6 weeks per commission.
Pricing anchored to a per-piece scale, not per-hour.

What's fading:

Mid-range editorial commissions ($150–$600) have been getting harder to fill at sustainable rates as clients can substitute machine-generated images.
Generalist freelance marketplaces (Upwork-style) are no longer viable for Maya's work tier.
Newsletters/Substack-style outlets often don't budget for commissioned illustration at all.

H1 isn't gone. It's narrowing. The high end (named, art-directed, signature-driven) is durable; the mid is the part that's being squeezed.

H3 — The pattern that's emerging (the fringe glimmers)

What's visible in fringe signals that hints at where the system might be moving:

A re-mattering of physical objects — riso prints, zines, hand-bound editions, screen-printed merch — as the commercial form for illustration that AI can't copy at near-zero cost. (Maya already does some of this; the question is whether it scales as a primary practice.)
Illustrator-as-curator-and-collaborator with AI tools — using image models as a sketching / iteration substrate while keeping a recognisable hand on the final piece. The artistic identity is the editorial filter.
Workshops, teaching, mentoring as a significant revenue line, not a sideline. The skill being taught is increasingly taste and editing rather than execution.
Direct patron relationships (Patreon-adjacent, paid newsletter illustration, subscription print drops) replacing some commissioned editorial work.
Cross-domain work — book covers + game key art + brand identity for restaurants + tarot deck — replacing single-domain specialisation. The signature travels.

H3 isn't a forecast. It's the set of practices already growing, in the gaps where H1 is failing, that could plausibly be how this kind of work is in 7 years.

H2 — The disruptive middle (and where Maya is right now)

H2 is where the contested signals live. Many of them point at the same activity but get read differently:

AI-integrated illustration practice. Some illustrators use Midjourney as a reference / moodboard tool, then draw by hand. Some use it for backgrounds and draw the figures. Some never touch it. Each move has a different professional reputation cost. Maya is in this discomfort.
AI-assisted art direction. Clients use AI to generate a brief faster, sometimes producing reference images they want Maya to "interpret" or "redo properly." This is destabilising the brief-giving relationship.
Hybrid teaching practices. Workshops that teach both hand technique and AI-tool fluency are starting to outperform either-pure variant in enrolment.
Authorship and disclosure norms in flux. Whether a piece "counts" as a Maya piece if it used AI for one step is a question with no settled answer in 2026.
Compensation for training data. Class actions against AI companies for unauthorised training — and proposed opt-out registries — are H2 in the strict sense: some of these reinforce H1 (artist guilds reasserting rights), some accelerate H3 (new compensation models emerging).

The standard Three Horizons move is to flag which H2 items are H2− (captured by H1, extending its life) vs H2+ (genuinely disruptive, opening H3). (3)

Adobe Firefly inside Photoshop: H2− — it extends the H1 toolchain and keeps Maya inside the Adobe orbit. Doesn't change what she does, only how fast she does some of it.
Maya's workshop teaching: H2+ — it points at a different revenue shape (subscription / cohort / live), not just a different tool.
Direct-patron monthly print drops: H2+ — a genuinely different relationship to audience.
Disclosure-required AI use on a commission: H2− if the disclosure is treated as a tax, H2+ if it changes what the work is.

The H2 question for Maya is not "should I use AI?" The H2 question is: which of the moves available to her are H2+ (open H3) and which are H2− (extend a fading H1)?

Try it yourself

Draw three rough S-curves on a sheet. Above each, name:
- H1 — today's pattern that's fading for your navigator: 5 specific
  things that have been working and are showing strain
- H3 — fringe signals of what might be growing: 5 practices visible
  on the edges that hint at a different way the work could be
- H2 — the disruptive middle: 5 activities your navigator is doing or
  could do that sit in flux

For each H2 item, label it H2− (extends H1) or H2+ (opens H3). Notice
which ones look one way to you in the morning and the other way at
night.

STEP 4 of 6 · HEXAGON 1 · SENSE-MAKE · Futures Wheel from one H2+

Trace one possible move outward — the real payoff (an audience, a voice) sits a ring or two out.

Three Horizons surfaces categories. The Futures Wheel chases consequences. We pick one of Maya's H2+ moves and trace it outward, so the picture acquires texture.

Driver (centre of wheel): Maya shifts her work mix over the next 3 years so that workshop teaching grows from 6 hrs/week to ~40% of her annual income.

This isn't a recommendation; it's a what-if. We're tracing what would happen if she made this move, so she can see what it pulls in.

First-order consequences

F1. Income shape changes from per-commission to subscription / cohort-based. Quarterly cash flow becomes more predictable.
F2. Her professional identity shifts. "Illustrator who teaches" becomes "illustrator-educator." Some of her referrals from editorial work dry up; others move sideways.
F3. Time use rebalances. Less drawing, more curriculum design, recording, marking, marketing the workshops.
F4. She acquires a small audience that follows her work and her teaching together. A list, a small Discord, monthly drops.
F5. Her relationship to AI tools changes. Teaching pushes her to articulate AI-use stances — required or forbidden, integrated or not.

Second-order consequences (tracing F4 — audience)

F4 → S1. Her income becomes more sensitive to platform risk (Substack, Discord, payment processor in Mexico). One platform decision can break a quarter.
F4 → S2. She starts producing publishable artefacts (print drops, workshop zines, recorded sessions) that have a longer commercial tail than commissions.
F4 → S3. She develops a voice in writing alongside drawing. Her newsletter copy starts driving as much demand as her images do. She didn't expect to become a writer.
F4 → S4. A new kind of friend group appears — other small-audience teaching illustrators, mostly outside Mexico. The local studio community recedes slightly.

Third-order consequences (tracing S3 — voice)

S3 → T1. Editorial commissioners begin to ask her for essay-and-illustration pieces rather than just illustration. Pay is better; relationships are different.
S3 → T2. Her teaching becomes a practice (sustained, opinionated, recognisable) rather than a side income. She starts being invited to speak.
S3 → T3. She begins to see herself as running a small business in a way she didn't at 32. Accounting, contracts, scheduling all change. She hires part-time help.
S3 → T4. A counter-move surfaces: missing the unstructured time of pure illustration, she cuts the teaching back from 40% to ~25% in year 5. The first move was a learning move, not a destination.

What the wheel surfaces

The audience branch is where the move actually pays off. The teaching is the vehicle; the audience is the value. (This is non-obvious from the first ring.)
Voice becomes load-bearing. Maya didn't plan to write more — the wheel surfaces that the move pulls in a writing practice she'd otherwise have ignored.
Platform risk is a real downside not visible from the first ring.
The third-order T4 — partial retreat — is important. It models navigation, not commitment. The move doesn't have to be permanent.

Try it yourself

Pick one H2+ move from your navigator's Three Horizons. Put it at the
centre of a Futures Wheel.
- Ring 1: 5 first-order consequences (direct effects on income, time,
  identity, relationships)
- Ring 2: for 2 of those, 3 second-order consequences
- Ring 3: for 1 second-order, 3 third-order consequences

End with: what does the wheel surface that wasn't in the move itself?
What's the *real* value being chased, that the move is just a vehicle
for? What's the downside ring you hadn't noticed?

STEP 5 of 6 · HEXAGON 1 · REFLECT · The navigation question, not the answer

The whole exercise has not produced a plan for Maya. It has produced attention — a structured way of seeing the next 7 years that is more textured than "AI is coming for illustrators" and more responsive than "I'll figure it out."

A reflective frame for Maya (and for anyone running this on themselves):

1. What did this surface that you weren't watching?

The voice / writing branch — not on the original radar.
Cohort effect of being 34: not too late to retool, not senior enough to coast.
Disclosure norms — a small thing now, plausibly a big determinant of what counts as a Maya piece.

2. What is genuinely H1 fading vs H3 emerging in your work, today? Honest naming. Some practices are not "core" — they're holdovers we keep up because they used to work.

3. Where might you be using H1 language for an H2+ move? "I teach a workshop on the side" is H1 language. "I run a teaching practice that funds my illustration work" is H2+ language. Naming changes what you protect.

4. What's an emergent move, not a planned one? Plans are H1-shaped: they assume the same world tomorrow. Emergence is what happens when you take a step and see what shows up. Three Horizons doesn't tell you to plan; it tells you to notice which steps are available.

5. What does this exercise refuse to answer?

Whether Maya should use AI tools. (Personal, value-laden, contextual.)
Whether she should move countries / cities. (Same.)
Whether the editorial industry survives. (Above her pay grade; not her unit of decision.)
How much money she'll earn in 7 years. (The wheel doesn't forecast.)

Refusing to answer these is part of the discipline. The tool is for seeing, not for deciding. The deciding is Maya's.

Try it yourself

Reflect on your own Three Horizons + Futures Wheel run, in 60 words or
fewer per question:
1. What did this surface that I wasn't watching?
2. What's H1 fading vs H3 emerging in *my* work today, honestly?
3. Where am I using H1 language for an H2+ move?
4. What's an emergent move, not a planned one, that's already available?
5. What does this refuse to answer — and is that OK?

STEP 6 of 6 · HEXAGON 1 · BRIDGE · One sentence to next quarter

Closing one-line discipline: turn the work into one attentional sentence for next quarter.

For Maya: "Over the next 3 months I will pay attention to the audience side of my teaching — newsletter list, workshop completion rates, any sign that voice matters more than I thought — and I will treat each AI-disclosure conversation with a client as data, not a tax."

Note what this sentence is not. It's not a plan to grow workshops to 40% of income. It's not a vow to use or refuse AI. It's an attention commitment, calibrated to surface signal. In 3 months, Maya re-runs the wheel.

That's individual navigation. The future doesn't get planned; it gets attended to. Three Horizons + Futures Wheel are tools for keeping the attention well-shaped.

What this example does and doesn't claim

Documented (with citations):

17% drop in image-creation freelance job posts after major image-model releases (1).
Bifurcation of the freelance design market and the 40–60% rate premium for AI-adopting designers, per Upwork data via industry coverage (2).
The Three Horizons framework's structure and the H2−/H2+ distinction (3).

Constructed:

Maya. Her income shape, her clients, her studio share, her cat — all composite illustration, not a real person. The Mexican peso figures and Roma Sur context are realistic but illustrative.
The Futures Wheel branches in Step 4 are speculative traces, not predictions about anyone's career.
The "third-order T4 — partial retreat" outcome is a deliberately constructed twist to model navigation rather than commitment.

Out of scope:

Sector-level labour market modelling. We chose individual scale on purpose.
Geographic comparative — Maya in Berlin would face a different scan; Maya in Lagos different again. A real navigation example for any specific person would not generalise from this.
AI ethics arguments. We name disclosure norms as a signal; we don't take a position.

References

[1] Hampole, M., Hsiao, A., Klenow, P. J., Lashkari, D., Mantegazza, P., & Tertilt, M. (2024). "Is generative AI a job killer? Evidence from the freelance market." Brookings Institution. brookings.edu. (Source for 17% decrease in image-creation freelance job posts after image-model releases.)

[2] Industry coverage of the freelance design market bifurcation includes Upwork's published research and industry summaries — see We And The Color, "Freelance Designers Can't Compete with a $20/Month AI Subscription — Here's What Actually Works Now" (weandthecolor.com); and 2727 Coworking, "AI's Impact on Freelancers: Job Trends, Skills & Outlook" (2727coworking.com). Specific income-shift figures (40–60% hourly premium for AI-adopting freelancers) trace back to Upwork's own data releases.

[3] Sharpe, B. (2013). Three Horizons: The Patterning of Hope. Triarchy Press. Framework developed with the International Futures Forum (IFF) and Andrew Curry. The H2−/H2+ distinction comes from Curry & Hodgson's later articulation of the framework. See also H3Uni Resource Library and Three Horizons Framework — NPC.

Methodological references

Glenn, J. C. (1972). "Futurizing Teaching vs. Futures Course." Social Science Record, 9(3), 26–29. (Origin of the Futures Wheel.)
Curry, A. & Hodgson, A. (2008). "Seeing in Multiple Horizons: Connecting Futures to Strategy." Journal of Futures Studies, 13(1), 1–20. (H2−/H2+ articulation.)