H1
The managed present
Zero to eighteen months. What already exists and is being defended. Evidence standard: measurable. This is where your operating plan lives, and where futures work adds least.
Approach · How the work runs
Three horizons.
Six lenses.
Five movements.
Foresight fails when it is a mood. Ours is a procedure — repeatable, auditable, and hostile to its own conclusions.
01 — Horizons
Hold all
three at once
Most strategy work collapses into one horizon and calls it realism. The discipline is keeping three live simultaneously, with different evidence standards for each.
H2
The contested middle
Two to six years. Where the fight actually happens — incumbents adapting, entrants scaling, regulation catching up. Evidence standard: directional and disputed. This is our centre of gravity.
H3
The emerging pattern
Seven years and beyond. Currently fringe practice, unfunded research, weird hobbyists. Evidence standard: coherence and plausibility. Cheap to watch, ruinous to ignore.
02 — Lenses
Six lenses,
no shortcuts
Every significant question is run through all six before synthesis. Each lens must produce at least one non-obvious insight or the pass does not count. No output ships with fewer than four applied.
01
First-principles
Strip the convention and the received wisdom. What does physics, thermodynamics, unit economics or human biology actually permit here — independent of what the industry currently believes?
02
Historical
Which analogous transition has already run to completion elsewhere? Electrification, containerisation, the browser wars. History does not repeat, but the failure modes are reusable.
03
Contrarian
Steelman the opposite of the consensus until it is genuinely uncomfortable. If we cannot argue the inverse well, we do not understand the position we are recommending.
04
Technical
What does the implementation actually require — in silicon, in kilowatts, in trained people, in latency budget? Most futures die on a constraint nobody costed.
05
Economic
Who captures the value, and what has to be true for them to keep capturing it? Follow the incentive structure, the margin pool and the subsidy. Business models decide which technology wins.
06
Geopolitical
Export controls, procurement rules, standards bodies, sovereign capability, cross-border data. Policy is a technology roadmap written by people who are not in the room with you.
03 — Movements
What twelve
weeks looks like
A full engagement runs roughly twelve to twenty weeks. Scanning retainers run continuously alongside. Every movement produces an artefact that survives the engagement.
Wk 1–4
Scan
Structured collection across the domain and two adjacents. Preprints, patents, capital flows, job ads, regulatory dockets, fringe practitioner communities. Everything sourced, dated and scored.
Wk 4–6
Diverge
Six-lens pass, run as separate analyses before any synthesis. Deliberately unreconciled at this stage — the disagreement between lenses is the raw material for the scenarios.
Wk 6–9
Construct
Two critical uncertainties become four worlds. Each gets its own economics, winners, losers, and the three indicators that would show it gaining. Reviewed adversarially before release.
Wk 9–18
Build
One scenario is chosen for a provocation prototype. Working software, tested with the people who would actually use it, instrumented to produce evidence rather than applause.
Wk 18+
Transfer
Cadence, dashboard, manual and source code land with your team. Six months of coached operation if you want it. The measure of success is that you stop needing us.
04 — Standing rules
Rules we
do not bend
Written down because they are easiest to abandon exactly when they matter.
Falsifiable or unpublished
Every claim carries something that would prove it wrong and a date by which we would expect to see it. A trend statement with no failure condition is decoration.
Source or silence
Assertions travel with provenance. If a number cannot be traced to a document, a dataset or a named interview, it does not enter the ledger.
Keep the wrong calls
We maintain a register of signals we misread, visible to the client. A foresight practice that never publishes its misses is selling comfort.
Build before you argue
When a debate runs past two sessions, it is cheaper to prototype the disputed thing than to keep meeting about it.
No single future
We will not reduce a scenario set to a base case on request. The set is the product; collapsing it destroys the option value you paid for.
Leave the soil better
Nitrogen fixing, not extraction. Every engagement transfers method, not just findings, and we say so in the statement of work.
Next
See what the method produces
Engagement shapes, artefacts, and three pieces of fieldwork we are allowed to describe.
Fieldwork