The Opportunity Graph™
A Living Map of Every Business Problem, Assumption, Opportunity, and Outcome
Most organizations have their data in silos. The Opportunity Graph™ connects everything — business problems, the assumptions that constrain them, the opportunities that dissolve those constraints, the outcomes achieved, and the verified impacts measured — into a single, traversable, continuously learning knowledge structure.
It is the structural memory of the Opportunity Intelligence™ platform. Every discovery, every validation, every capture, and every outcome makes the graph smarter. The graph is the moat.
The Seven Node Types That Structure Organizational Knowledge
Every entity in the Opportunity Graph™ is one of seven node types. Each has a defined schema, a confidence score, and a set of valid inbound and outbound edge types. This schema is the graph's grammar — it ensures every connection is meaningful and queryable.
A specific, named business problem that is costing the organization money, time, or opportunity. Problems are the anchor nodes of the graph — every other node type connects back to a problem.
INBOUND: DETECTED_BY (Signal), MANIFESTS_AS (Symptom) | OUTBOUND: CONSTRAINED_BY (Assumption), GENERATES (Opportunity), RESOLVED_BY (Outcome)
A belief the organization holds that constrains decision-making. Assumptions are the most dangerous nodes in the graph — when invalidated, they unlock entire subgraphs of new opportunities.
INBOUND: CONSTRAINS (BusinessProblem) | OUTBOUND: CREATES (FalseDilemma), LIMITS (Opportunity), SUPPORTED_BY (Evidence)
A specific, dollar-quantified opportunity to create or recover value. Opportunities are the graph's value-bearing nodes — the entire graph exists to generate, validate, and track them.
INBOUND: GENERATED_BY (BusinessProblem), UNLOCKED_BY (Assumption invalidation), ENABLED_BY (FalseDilemma resolution) | OUTBOUND: HAS_PATH (CapturePath), PRODUCES (Outcome), VERIFIED_BY (Verification)
A specific strategy for capturing an opportunity. Each opportunity can have multiple capture paths — the graph models each as a separate node so they can be compared, simulated, and selected independently.
INBOUND: CAPTURES (Opportunity) | OUTBOUND: REQUIRES (Resource), HAS_MILESTONE (Milestone), PRODUCES (Outcome)
A measured result — the actual value realized (or not realized) from capturing an opportunity. Outcomes close the learning loop: they teach the graph which opportunity types, capture paths, and assumptions produced real value.
INBOUND: PRODUCED_BY (CapturePath), RESOLVES (BusinessProblem) | OUTBOUND: ATTRIBUTED_TO (Impact), UPDATES (Assumption confidence)
A statistically verified impact — the gold-standard node. Verified Impacts are what separate the Opportunity Graph™ from a mere database. They are auditable, statistically rigorous measurements that the graph uses to calibrate future predictions.
INBOUND: DERIVED_FROM (Outcome) | OUTBOUND: CALIBRATES (Opportunity scoring), STRENGTHENS (Confidence), PROVES (Assumption validity)
Raw or processed data ingested from the organization's signal environment. Signals are the leaf nodes — they feed the graph with new information. Most signals are transient; only those that trigger opportunity detection persist as connected nodes.
OUTBOUND: DETECTS (BusinessProblem), INDICATES (Opportunity), CORROBORATES (Assumption), CONTRADICTS (Assumption)
The Twelve Edge Types That Define Valid Graph Traversal
Edges are not generic "related to" links — each edge type has a defined semantic meaning, a direction, a confidence weight, and validation rules. This strict edge grammar is what makes the graph queryable with precision and enables automated traversal for opportunity discovery.
The assumption limits how the organization can address the problem. Invalidation of the assumption removes the constraint.
The problem creates a specific opportunity to recover or create value — the problem is the 'why' behind the opportunity.
Invalidating this assumption unlocks an opportunity that was previously hidden by the false constraint.
The assumption is the root cause of a false dilemma — the dilemma exists because the assumption is treated as fact.
This specific capture strategy is a viable approach to converting the opportunity into value.
Executing this capture path produced (or is expected to produce) this specific outcome.
Statistical analysis of this outcome produced a verified, auditable impact measurement.
The realized outcome addressed (fully or partially) the original business problem.
Verified impact data recalibrates the scoring models — ODS, BIIQ, OVS — making future predictions more accurate.
Signal analysis (Scout™) detected a previously unknown business problem.
New signal data supports the continued validity of an assumption.
New signal data challenges an assumption — this is the highest-value edge type. Contradiction triggers re-evaluation of every node downstream.
Edge Validation Rules
Every edge created in the graph MUST pass validation. These rules prevent the graph from accumulating meaningless or incorrect connections.
How the Graph Discovers What You Don't Know to Look For
The graph's value is not just in storing connections — it's in automated traversal that surfaces insights no human would think to query. The traversal engine walks the graph according to defined patterns, scoring each path for relevance and novelty.
Assumption Cascade Traversal
A CONTRADICTS edge appears on an Assumption node
BFS from the contradicted assumption, following: Assumption → CREATES → FalseDilemma, Assumption → LIMITS → Opportunity, Opportunity → HAS_PATH → CapturePath. Score each downstream node by how much its expected value changes when the assumption is invalidated.
Assumption Impact Cascade Report — every downstream opportunity and capture path affected by the new evidence.
Opportunity Pattern Matching
A new Opportunity node is created
Extract the opportunity's feature vector (type, domain, value, problem class). Cosine similarity search across all historical Opportunity nodes. Return top-K historical matches with their outcomes and verified impacts attached.
Pattern Match Report — 'This opportunity looks like 3 others that produced $X in verified impact. Here's what worked.'
Problem Cluster Detection
Continuous — runs every 24 hours
Community detection (Louvain algorithm) on the BusinessProblem → Opportunity subgraph. Find clusters of problems that co-occur across organizations, share common assumptions, or produce similar opportunity types. Surface clusters that are growing (new signals arriving) or unresolved (many open opportunities, few outcomes).
Problem Cluster Map — groups of related problems, ranked by aggregate TAI™ and resolution velocity.
Cross-Vertical Transfer
A VerifiedImpact node is created with high confidence
From the VerifiedImpact, traverse: Impact → DERIVED_FROM → Outcome → PRODUCED_BY → CapturePath → HAS_PATH → Opportunity → GENERATED_BY → Problem. Extract the pattern: Problem class + Assumption profile + Capture approach → Verified Impact magnitude. Search for structurally similar problem subgraphs in OTHER verticals. Score each match by structural similarity × cross-vertical applicability discount.
Cross-Vertical Opportunity Transfer Report — 'The capture approach that worked for behavioral health credentialing could apply to legal client intake. Estimated impact: $X.'
Assumption-Constraint Discovery
On Scout™ Assumption Detection run (Module 1)
For each newly detected Assumption, traverse: Assumption → CONSTRAINS → Problem → GENERATES → Opportunity. Score by: (Assumption certainty⁻¹ × Assumption impactIfWrong) + (Problem severity × number of constrained opportunities). Rank by compound score.
Highest-Leverage Assumptions Report — 'These 5 assumptions, if invalidated, would unlock $XM in constrained opportunity value.'
Impact Decay Monitoring
Continuous — runs every 6 hours
Monitor all Outcome nodes where: (now − realizedAt) > 90 days and no new VerifiedImpact has been DERIVED_FROM the Outcome. Also monitor Opportunity nodes in 'Captured' status with no linked Outcome. Flag for review — the graph may be missing measurement data.
Measurement Gap Report — opportunities and outcomes that lack verified impact data. Highlights where the learning loop is broken.
The Confidence System That Separates Signal from Noise
Not every node and edge in the graph is equally trustworthy. The Confidence Scoring System™ assigns a dynamic confidence score to every node and edge, continuously updated as new evidence arrives. This is what prevents the graph from becoming a garbage-in-garbage-out machine.
Node Confidence Score (NCS™)
Graded by evidence type: Verified Impact (+50), Causal Analysis (+40), A/B Test (+35), Cohort Study (+25), Expert Consensus (+15), Anecdotal (+5). Highest grade wins.
Number and weight of inbound edges. A node with 10 CORROBORATES edges from diverse signal sources is more trustworthy than one with 1.
Exponential decay function: 100 at creation, halves every 180 days without new corroborating evidence. Nodes with recent activity maintain high recency.
Number of independent signal sources that support this node's existence. Cross-source agreement is the strongest confidence signal.
Edge Confidence Score (ECS™)
Edges with ECS below 0.25 after 365 days are automatically pruned. CONTRADICTS edges are exempt — they persist even at low confidence because a single weak contradiction signal can be the seed of a major discovery. The graph grows through connection and stays healthy through selective forgetting.
How the Graph Discovers Opportunities You Never Knew Existed
The graph doesn't just store what you know — it discovers what you don't. These six discovery mechanisms run continuously, surfacing opportunities that no human analyst would find through traditional query-based approaches.
Structural Hole Discovery
In social network theory, structural holes are gaps between clusters. In the Opportunity Graph™, a structural hole is a missing connection where a connection should exist — and that absence IS the opportunity.
Run community detection on the BusinessProblem → Opportunity subgraph. Identify pairs of Problem clusters that share similar Assumption profiles and Opportunity types but have NO cross-cluster edges. The missing PRODUCES edges between one cluster's CapturePaths and the other cluster's Outcomes indicate unapplied solutions.
Cluster A (home healthcare) has CapturePaths that successfully resolved scheduling problems. Cluster B (behavioral health) has structurally similar scheduling problems with NO linked CapturePaths. The graph discovers: apply Cluster A's capture approach to Cluster B's problems.
Assumption Invalidation Cascade
When an Assumption node flips from Validated to Invalidated, it doesn't just affect the problems it directly constrained. The graph traces the full downstream cascade to identify every Opportunity that was previously hidden by that now-invalid assumption.
Dijkstra-style weighted traversal from the invalidated Assumption node. Edge weights are inverted (lower = more affected). Traverse Assumption → CONSTRAINS → Problem → GENERATES → Opportunity. Each hop multiplies the impact factor. Generate new Candidate Opportunity nodes for every downstream node that was status=Hidden or had severely discounted ODS scores.
Assumption 'enterprise customers require on-premise deployment' is invalidated by market data. Cascade reveals: 47 previously hidden opportunities across 12 problem domains, total estimated value $18.3M.
Anomalous Path Detection
Most paths through the graph follow predictable patterns. Anomalous paths — where the sequence of node types or edge types deviates from the norm — often indicate overlooked opportunities or flawed mental models.
Train a graph neural network (GNN) on the normalized path distribution of all successful Opportunity → CapturePath → Outcome → VerifiedImpact paths. Flag paths whose embedding distance from the nearest successful path exceeds 2σ. Anomalous paths are surfaced for expert review — they're either errors (and should be pruned) or genius (and should be pursued).
A path where an Opportunity connects directly to VerifiedImpact (skipping CapturePath and Outcome) is anomalous — it means value was realized but the capture mechanism was undocumented. The graph flags this as both a documentation gap AND a potentially novel capture pattern.
Compound Opportunity Synthesis
Some opportunities only exist when two or more others are captured together. The graph identifies these compound opportunities by detecting sets of Opportunity nodes whose combined capture (via connected CapturePaths) unlocks additional value not accounted for in either individually.
Frequent subgraph mining across the Opportunity → HAS_PATH → CapturePath subgraph. Identify pairs (and triples) of Opportunities whose CapturePaths share ≥ 70% of their resource requirements. Merge their Impact Simulations to model the compound effect. If compound TAI > Σ(individual TAI) by > 15%, surface as a Compound Opportunity.
Opportunity A (automate patient intake) + Opportunity B (integrate scheduling API) have 80% overlapping resource requirements. Captured together, they unlock Opportunity C (real-time capacity optimization) which neither enabled alone. Compound TAI: $4.7M vs. $3.1M individually.
Temporal Pattern Recognition
Opportunities have lifecycles. The graph learns the temporal signatures of different opportunity types — when they emerge, how long they persist, and what signals precede them — and uses those signatures to predict opportunities before they fully materialize.
Time-series analysis on Signal → DETECTS → Problem → GENERATES → Opportunity paths. For each opportunity type, identify the leading indicator signal pattern that preceded detection with highest precision. Deploy pattern watchers that trigger pre-opportunity alerts when the leading indicator pattern appears.
Historical analysis shows that 'underbilling opportunity' is preceded 45–60 days by a pattern of: increasing support ticket volume + declining contract amendment frequency + above-average NPS. Pattern watcher triggers when all 3 signals co-occur.
Counterfactual Opportunity Generation
What would the graph look like if a key assumption were false? What if a competitor captured this opportunity first? Counterfactual reasoning generates hypothetical subgraphs, scores them, and surfaces the most valuable 'what if' scenarios as candidate strategies.
For each high-impact Assumption node (ADS ≥ 70), generate a counterfactual subgraph where the assumption is false. Traverse the counterfactual subgraph to enumerate all newly possible Opportunities. Score with Counterfactual BIIQ™ (cBIIQ): BIIQ × P(Assumption is actually false). Rank and surface top-N.
Counterfactual: 'What if our largest competitor exits market segment X?' Generates 23 new opportunity candidates, top-3 cBIIQ scores: 87, 82, 79. Estimated total counterfactual TAI: $31.2M.
Graph Discovery Performance Metrics
The Opportunity Graph™ Is the Structural Moat of the Platform
Every opportunity discovered by Scout™, every assumption validated by the Trust Engine™, every impact verified by IVS™ — feeds into the graph. The graph is the platform's memory, its learning engine, and its primary defensibility. See how the full platform architecture works together.