Overview

The Wicked-Future Analysis Framework (WFA) is the depth-molecular mode in T6 (future exploration) — the heaviest forward-looking analytical operation Ora currently builds. It runs when a forward-looking question resists single-method analysis: scenarios alone do not carry calibrated probabilities, probability-forecasting alone misses the divergence narratives, pre-mortem alone evaluates one plan rather than the broader future. WFA’s value is in the integration. Scenarios with probability bands. Scenarios stress-tested against pre-mortem failure pathways. Divergence points to monitor that serve double duty as scenario-distinguishers and failure-mode leading indicators. The integrated forward architecture is the dialectical product of three complementary analyses; no single component could produce it alone.

The framework runs three component modes (scenario-planning for narrative scenarios, probabilistic-forecasting for calibrated probability bands, pre-mortem-action for adversarial-future stress-testing) and three synthesis stages (scenario-probability overlay, failure-pathway stress test, integrated future architecture). A fourth component — backcasting for the constructive-future stance (working backward from a desired future) — is gap-deferred per CR-6, and the framework documents the deferral explicitly rather than substituting. The gap-flag is mandatory and visible in the final artifact, not buried in the confidence map.

The framework’s load-bearing intellectual content is the silo-aggregation guard, the no-most-likely discipline, the pre-mortem-against-leading-scenarios requirement, and the silent-gap refusal. The silo-aggregation guard says synthesis must integrate rather than concatenate the three components — outputs presenting scenarios, probabilities, and failure pathways as three side-by-side reports without producing claims no single component could have produced are silo-aggregation failures. The no-most-likely discipline preserves scenario planning’s anti-prediction stance: even when probability bands cluster, the framework refuses to designate one as most likely; the user is forced to monitor divergence points rather than commit prematurely. The pre-mortem-against-leading-scenarios requirement applies the failure-pathway stress test to the highest-probability scenarios (the ones the user is most likely to plan for). The silent-gap refusal is the constructive-future deferral handling: the gap flag is visible and explicit; backcasting is not silently substituted.

The framework deliberately resists four patterns: trend extrapolation bias (counteracted by explicit critical-uncertainty axes); silo aggregation (integration requirement at each synthesis stage); pre-mortem omission (pre-mortem must run against leading scenarios, not just intervention candidates); silent gap (mandatory visible gap-flag for deferred backcasting).

The framework answers questions like: What will the AI agent ecosystem look like in five years? What could disrupt our supply chain over three years? Give me probability-weighted futures plus their failure pathways plus what to monitor — not three separate reports.

Systemic context

Wicked-Future Analysis is the depth-molecular operation in T6 (future exploration), composing siblings into an integrated forward analysis. The framework sits in T6’s depth ladder above consequences-and-sequel (T6-light, atomic, forward projection), scenario-planning (T6-thorough, atomic, narrative-output), probabilistic-forecasting (T6-thorough, atomic, probability-output), and pre-mortem-action (T6 stance-counterpart, atomic, adversarial-future). The territory framework is Framework — Future Exploration. WFA is the heaviest analytical mode in T6 currently buildable; the constructive-future stance (backcasting) is gap-deferred per CR-6 and the framework documents the deferred-component handling explicitly rather than substituting. WFA invokes from three paths: user invocation directly with a forward question; pipeline-dispatched from the four-stage pre-routing pipeline classifying the prompt as T6-future-exploration, depth-molecular; and handoff from another mode (scenario-planning or probabilistic-forecasting has surfaced that the question warrants integrated treatment). WFA shares the multi-future treatment with sibling Wave 4 modes — decision-architecture (T3, uses pre-mortem-action as Stage 4) and wicked-problems (T2, uses scenario-planning at Stage 3) — and is paired with [[Framework — Worldview Cartography Analysis]] as the other T6/T9 depth-molecular operations under the same architectural pattern.

Ora’s output

Example question: What does the agent ecosystem look like in 5 years? I keep getting fragmentary takes — labs racing to capability, regulators racing to governance, infra costs unclear. I want a real architecture I could use to plan a multi-year product strategy.

Ora’s response (running WFA on the named question):

The question is genuinely tangled, the time horizon is long enough that single-method analysis is insufficient, and the user wants integrated output rather than three separate reads. WFA is the right framework. I’ll run the three components and three synthesis stages, with the constructive-future gap-flag visible at the end.

Stage 1 — Scenario Planning. Driving forces: predetermined (rising compute investment, growing developer interest, capability scaling within current paradigms); critical uncertainties (regulatory regime — laissez-faire vs. licensing-required; capability trajectory — plateau vs. continued gains). The axes are independent (regulatory choice depends on political coalitions and incident-driven reaction; capability depends on methods research and data availability). Four scenarios: Licensed Ecosystem (high reg, high capability) — incumbents licensed, infrastructure consolidated under top firms. Open Cambrian (low reg, high capability) — exuberant fragmentation across thousands of agent products. Bureaucratic Stasis (high reg, plateau) — regulation locks in current architectures. Diffusion at Scale (low reg, plateau) — agent products spread into existing workflows. Wild card: “Capability cliff event” — a major incident triggers emergency global regulation within 18 months regardless of trajectory. No scenario designated “most likely.”

Stage 2 — Probabilistic Forecasting. Per scenario with reference class, base rate, inside-view adjustment, probability range: Licensed Ecosystem 0.30-0.45 (reference: past licensed-tech regimes; base ~40%). Open Cambrian 0.25-0.45 (reference: post-tech-revolution fragmented ecosystems; base ~35%). Bureaucratic Stasis 0.10-0.25 (rare combination; base ~15%). Diffusion at Scale 0.20-0.35 (base ~30%). Cross-scenario forecast: “Major agent-driven incident triggering emergency regulation” 0.20-0.35.

Synthesis Stage 1 — Scenario-Probability Overlay. Probability bands cluster (Licensed Ecosystem and Open Cambrian both 0.30+); the framework refuses to designate either most likely. Divergence points: 6-12 months (EU AI Act enforcement actions); 12-18 months (capability benchmarks); 18-24 months (top-5 inference-compute concentration ratio). Coherence check passes.

Stage 3 — Pre-Mortem Per Scenario. Licensed Ecosystem failure — compliance burden squashed innovation; capability advanced overseas in unlicensed jurisdictions; F1 (assumption that licensing would be calibrated; was punitive). Open Cambrian failure — fragmented ecosystem produced repeated catastrophic incidents; public backlash triggered draconian retroactive regulation; F2 (interaction: assumed self-regulation; correlated incidents overwhelmed it). Bureaucratic Stasis failure — capability plateaued in mainstream paradigms but breakthrough emerged in unlicensed jurisdictions; F3 (context-shift). Diffusion at Scale failure — value capture concentrated in software incumbents that had distribution; F4 (motivational: capability assumed decisive; was distribution).

Synthesis Stage 2 — Failure-Pathway Stress Test. Open Cambrian’s F2 (catastrophic incident → backlash → draconian regulation) is the highest-impact failure pathway and is also the trigger for the Stage 1 wild card. The wild card is structurally a failure mode of Open Cambrian, not an independent fifth scenario — Stage 3 revealed this; Stage 1’s wild-card framing did not. Divergence-points-serving-double-duty: “EU AI Act enforcement actions in first 12 months” distinguishes regulatory axis AND leads F1; “Agent-driven incident damage in 2026-2027” distinguishes Open Cambrian survival AND leads F2 → wild card.

Synthesis Stage 3 — Integrated Future Architecture. Three claims no single component could have produced: (1) the wild card is structurally a failure pathway of one matrix scenario, not an independent fifth — pre-mortem revealed this. (2) Open Cambrian and Licensed Ecosystem have overlapping probability bands but radically different preparation postures; recommendation is reliability tooling (robust across both) and divergence-point monitoring rather than betting on either. (3) Bureaucratic Stasis, despite being lowest-probability, has the most asymmetric preparation cost — easy to prepare for, expensive to be caught unprepared in; probability formalism alone would have de-prioritized it.

Constructive-future gap-flag (mandatory, visible). This analysis does NOT include backcasting. The backcasting mode is gap-deferred per CR-6. WFA composes around its absence by anchoring scenario-planning, probabilistic-forecasting, and pre-mortem-action. The constructive-future stance — working backward from a desired future to identify the chain of events that would produce it — is not substituted. Consumers requiring constructive-future framing should compose WFA with downstream goal-articulation work or wait for the eventual backcasting mode build.

That is what WFA produces. The user now has an integrated forward architecture: four scenarios with probability bands, one wild card now reclassified as a failure pathway of one matrix scenario, three divergence points with monitoring priorities, four pre-mortem failure pathways with leading indicators, three load-bearing claims the dialectical synthesis produced that no component could have produced alone, and an explicit visible gap-flag noting what the analysis does not do. The user can plan a multi-year strategy with explicit awareness of which futures are plausible, which preparation moves are robust across futures, which moves are scenario-dependent, and what signals to monitor for early evidence of which scenario is materializing.

Commercial AI comparison

Comparison content auto-populates when the comparison-refresh framework runs against this question. Drafters do not author this section.

Brief comparison commentary

Auto-populates with the comparison content above.

How to use this framework

You can run the Wicked-Future Analysis pattern with any AI of your choice. The composition is multi-pass against the three-component, three-synthesis-stage protocol.

The prompt:

[Paste the framework specification.]

Run WFA on this forward question.

Forward question: [The future-shaped question being explored.]

Time horizon: [Years out.]

Key uncertainties (optional): [If you have identified driving forces or critical uncertainties, share them.]

Prior scenarios (optional): [If you have done preliminary scenario thinking.]

Prior forecasts (optional): [If you have probability estimates on key outcomes.]

Intervention candidates (optional): [Possible actions or strategies you’re considering — the framework will run pre-mortem against them in addition to scenarios.]

The AI runs Stage 1 (scenario planning), Stage 2 (probabilistic forecasting), Synthesis Stage 1 (scenario-probability overlay), Stage 3 (pre-mortem against scenarios and any intervention candidates), Synthesis Stage 2 (failure-pathway stress test), and Synthesis Stage 3 (integrated future architecture). The output follows the eight-section template: forward question and horizon; scenario set with probability bands; divergence points; failure-pathway stress test findings; integrated forward architecture; constructive-future gap-flag (mandatory); residual uncertainties; confidence map.

For best results:

  1. State the time horizon explicitly. Probability bands beyond five years widen substantially; the framework produces honest analyses but its calibration is much stronger at shorter horizons. If your horizon is ten or twenty years, expect wider ranges and explicit long-horizon humility.
  2. Provide intervention candidates if you have them. The pre-mortem stage stress-tests both scenarios and intervention candidates. Naming intervention candidates lets the framework produce action-oriented failure analysis alongside the scenario-failure analysis.
  3. Don’t suppress the gap-flag. The constructive-future deferral is a real architectural commitment; if your question requires backcasting, WFA tells you that explicitly rather than producing a degraded substitute. Pair WFA with downstream goal-articulation work or wait for the backcasting mode.
  4. Don’t ask WFA to designate a most-likely scenario. The no-most-likely discipline is load-bearing for the framework’s value. If the framework reports clustered probability bands, the right response is to monitor divergence points, not to ask the framework to break the tie.
  5. Use the divergence points actively. The highest-leverage signals are those that distinguish scenarios AND serve as leading indicators for failure modes. Build monitoring around them; revisit the architecture as signals come in.

The framework is deliberately tool-agnostic. The three-component composition, the synthesis-stage discipline, the no-most-likely rule, and the visible gap-flag are conceptual disciplines that survive the lift to any environment. The integrated forward architecture is plain markdown.

Other examples

  • A regulatory-shift forecast for a regulated industry — A decade-out forecast on how an upcoming regulatory shift reshapes the sector. Critical uncertainties: which regulatory model the shift converges on; international harmonization scope. Pre-mortem runs on each scenario plus the user’s lean-in vs. defensive-posture candidates. The integrated architecture identifies divergence points (specific rule-making milestones), the scenarios where defensive-posture has catastrophic failure pathways, and robust strategies (compliance-infrastructure investments paying off across all scenarios). Demonstrates integration producing preparation guidance no single component could.

  • A supply-chain disruption forecast for a global manufacturer — Three-year horizon. Uncertainties: geopolitical realignment scope; climate-disruption frequency. The integrated architecture finds that the highest-probability scenario is well covered by current resilience investments, while two lower-probability scenarios have catastrophic failure pathways under those same investments. Recommendation: maintain current posture but add contingent investments tied to early signals of either lower-probability scenario. Demonstrates failure-pathway stress test re-balancing priorities away from probability-only ranking.

  • An open-ended sector forecast for a venture investor — Consumer-AI ecosystem at five years to inform investment thesis. Pre-mortem runs on each scenario without intervention candidates (the question is about the landscape, not investor actions). The architecture identifies which scenarios the portfolio is positioned for, which are uncovered, and which carry portfolio-level risk concentration. The constructive-future gap-flag is invoked: choosing a target scenario and reverse-engineering the portfolio is backcasting work and not part of WFA’s output. Demonstrates gap-flag handling.

Citations

The Wicked-Future Analysis Framework draws on traditions in forecasting, scenario planning, and prospective-hindsight stress-testing. Wack’s “Scenarios: Uncharted Waters Ahead” (HBR, 1985) and Schwartz’s The Art of the Long View (1991) provide the scenario-planning discipline against single-future prediction. Tetlock and Gardner’s Superforecasting (2015) and the Good Judgment Project research provide the calibration discipline — base-rate-anchored reference classes, inside-vs-outside view separation, ranges rather than points. Klein’s “Performing a Project Premortem” (HBR, 2007) provides the pre-mortem prospective-hindsight stance. Taleb’s The Black Swan and Antifragile provide the wild-card framing and long-horizon humility. Knight’s Risk, Uncertainty, and Profit (1921) underpins probability bands as honestly wide rather than falsely precise.

The integration discipline (silo-aggregation guard, no-most-likely rule, pre-mortem-against-leading-scenarios requirement) is internal to Ora and operationalizes the commitment that depth-molecular operations must produce claims no atomic component could produce alone. The constructive-future deferral is documented per CR-6 — explicit handling of deferred components is itself a discipline against silent substitution.

The framework is single-author and originated 2026-05-01 from the T6 territory consolidation alongside T9’s worldview-cartography mode under the same architectural pattern.

Downloads

  • Framework specification (PDF) — link to ora-ai.org canonical artifact when published
  • Framework specification (plain text) — link to ora-ai.org canonical artifact when published
  • Full white paper (PDF) — link when published