HORIZON
OBSERVATION
01 / Frame

Who gains when
experience becomes
an AI asset?

AI can make professional work faster. Whether that creates new demand—and who captures the returns—remains open.

Follow reusable expertise through four markets, a professional firm and the employee who leaves.

Explore the drivers

One present. Several plausible futures.

The widening cone represents a range of possibilities, not a prediction.

Four possible AI futures within a cone of plausibilityThe cone widens from 2026 to 2035. Four points preview scenarios; position and width do not represent likelihood.

Four markets, different returns.

Select a point to preview a market.
The provocation

What if your firm could keep using parts of your experience after you leave?

Go to the employee experiment ↗
02 / Scan

Useful AI needs more
than better models.

Ten drivers shape whether expertise can be captured, trusted, delivered and sold.

The scan spans society, technology, the economy, the environment and policy.

Explore the forces

Select a dot or use the menu to inspect a driver.

Selected driver

Which of these drivers could lead to meaningfully different markets?

Why these two questions?
03 / Prioritize

Two questions change
who captures value.

Demand changes the size of the market. Access changes who can compete for it.

Axis A · New paid demand

Will AI create
new paying demand?

Lower costs may mainly replace existing spending—or make previously inaccessible services worth buying.

LimitedMostly familiar purchasesSubstantialNew buyers and services
Axis B · Customer access

Who can reach
those buyers?

A few gateways may control discovery and distribution—or many providers may have viable routes to customers.

ConcentratedFew practical gatewaysContestableViable alternatives

Why combine them?

Either access pattern can coexist with either level of demand. Crossing them separates market growth from bargaining power.

What about the other eight?

Reliability, trust, learning, cost, energy and policy remain conditions to test in every world.

Driver assessments

Ratings are qualitative judgments. The axes are chosen for the value-capture question, not automatically by rank.

Crossing the two axes gives us four possible markets.

Enter the four markets
04 / Scenarios

Four markets.
The same expertise.

A reusable method can be valuable in very different ways, depending on the market around it.

Your chosen market carries through the firm, employee and ownership chapters.

Read downward: more new paid demandRead rightward: more contestable access

Possible markets, not forecasts. Real markets may combine their features.

05 / ImplicationsCaptive Efficiency

What changes inside
a professional firm?

Market conditions change client behaviour—and the firm’s economics.

    Inside the firm

    Select clients, delivery, revenue or people to see what changes.

    Effects across the value chain

    Select a participant to see how it is affected. Lines show dependencies, not measured money flows.

    The firm needs a repeatable method.
    People’s work is one possible source.

    The incentive

    A possible response

    The raw material

    Now add one assumption: permitted work traces can help train or adapt a useful model.

    Follow the employee’s experience
    06 / AI artifactsCaptive Efficiency

    Can experience keep working
    after someone leaves?

    The same what-if applies in all four markets; it is not a fifth scenario.

    The limit of capture

    Some recorded patterns may transfer. A person’s full judgment, client context and responsibility do not automatically transfer with them.

    Scroll through work → capture → departure → reuse.

    Skip animation ↓

    The employee leaves.
    The recorded patterns remain.

    The model may continue to support new work. Current professionals still need to check its output and take responsibility.

    Who can keep using the asset, take it elsewhere, or share in what it earns?

    Keep the market. Change the terms. ↓
    07 / OwnershipCaptive Efficiency

    Who owns your
    AI artifacts?

    You leave. The model keeps working.
    What do you keep?

    Three possible agreements for the same AI artifact.

    All three agreements can exist within your selected market.

    Before your experience becomes an asset, decide your terms.

    Follow the argument

    Evidence & method

    We explore how reusable professional expertise could create value across four markets—and who captures the returns.

    How the reasoning is built

    1. Frame: define professional expertise, the wider AI value chain and a 2035 horizon.
    2. Scan: collect signals across social, technological, economic, environmental and political systems.
    3. Prioritize: judge impact and uncertainty, then choose two questions relevant to value capture.
    4. Scenarios: cross demand and access to explore four plausible markets.
    5. Implications: trace possible consequences for firms, customers and people.
    6. Stress-test: introduce the same employee-artifact assumption into each market.
    7. Options and monitoring: compare rules for reuse, prepare robust choices and identify evidence that could challenge the story.

    The method draws on the UK Government Office for Science Futures Toolkit and the OECD Strategic Foresight Toolkit. The scenarios and assessments are authored interpretations.

    What the diagrams do—and do not—claim

    The cone shows a range of possibilities; it does not assign likelihood. Radar positions group drivers by STEEP category. The scenario axes represent market conditions. Terrain, line thickness and animation do not represent measured money flows, market shares or probabilities.

    Different sectors and regions may combine features of the four worlds. Ownership arrangements are illustrative negotiated rules. Client permissions, confidential material, reliability and accountability remain separate constraints.

    Strategic options and signals to watch

    Prepare across futures

    • Make permissions, contribution and portability explicit before reuse.
    • Keep investing in people’s learning and the capacity to review unfamiliar cases.
    • Test a repeatable client outcome and a viable route to buyers before scaling the asset.

    Watch what would change the story

    • Demand: repeat purchases from previously unserved clients.
    • Access: switching costs and dependence on a few gateways.
    • Capability: performance on new cases and the effort required for review.

    In your selected world:

    Evidence baseline

    Research baseline: 12 October 2026. Historical evidence informs these scenarios; it does not prove them or establish that an employee’s expertise can be copied.

    What could challenge the central provocation?

    If work traces fail on new cases, need constant expert correction, or cannot be reused with appropriate permission, the artifact may remain a limited assistant. New demand may expand professional work alongside automation.