Industrial Genome Platform

Product Overview

Black-box system identification for industrial companies. Infers internal dynamics — constraints, bottlenecks, fragility — from observable public signals.

Owner: Ryan Cahalane · ryan.cahalane@lns-global.com Status: Early Build · April 2026 Stack: Python · FastAPI · Next.js · Supabase

The Problem

Industrial systems are opaque.

Companies operate as complex, dynamic systems — but the underlying constraints that drive performance are rarely visible until after they show up in results.

Financials are lagging indicators. By the time issues appear, the damage is already done.
Internal reporting is fragmented, slow, and often biased toward how the business should work, not how it actually behaves.
External stakeholders — operators, leaders, investors, partners — lack a consistent way to identify where constraint energy is building: in supply, demand, flow, or cost.
Identifying the true bottleneck still relies on interviews, site visits, and intuition — taking weeks or months while conditions continue to change.
There is no standard "fingerprint" for how an industrial system behaves under pressure — only anecdotes, snapshots, and delayed metrics.

As a result:

decisions are made too late · resources are deployed inefficiently · opportunities are missed or misdiagnosed

The problem isn't lack of data — it's lack of a system that turns observable signals into a coherent view of how the business is actually operating.

Core Concept

Treat industrial companies as dynamic systems that can be inferred — not just observed.

Instead of relying on internal access or delayed reporting, the platform uses observable signals to reconstruct how a business is behaving in real time.

External Inputs

Commodities · labor markets · demand signals · logistics · pricing

Behavioral Signals

Hiring patterns · content signals · operator chatter · activity patterns

Structural Inputs

Process type · industry dynamics · technology footprint

These are combined using a layered model:

Genome

System state

Identifies whether the system is resilient, constrained, fragile, oscillatory, saturated, or decoupled.

ECL

External Constraint Layer

Quantifies external pressures acting on the system — commodity cost, energy, labor, demand, logistics.

PFSL

Pre-Financial Signal Layer

Detects early behavioral signals before financial impact — inventory stress, margin compression, demand velocity.

Graph

Knowledge Graph

Connects entities, signals, constraint nodes, and outcomes across the industrial universe.

ENOX

Execution Layer

Translates insight into action via expert matching and execution pathways.

What That Enables

Early Detection

Identify constraints before they show up in financials — weeks or quarters ahead of the income statement.

Constraint Precision

Distinguish supply vs. demand vs. execution bottlenecks. Know where the energy is concentrated, not just that performance is off.

Inflection Detection

Detect early inflection points — improvement or deterioration — before they're visible in reported numbers.

Behavioral Comparison

Compare systems based on how they behave under pressure, not just sector or size. Peer benchmarking on system dynamics.

Signal to Action

Convert signals into action — investment opportunities, expert matches, and execution pathways.

Fingerprint dimensions

Archetypes

Constraint nodes (D1–X3)

16+

Industry signal bundles

System Architecture

Industrial Genome platform: signal acquisition → Genome model → intelligence hub → expert marketplace

Value Architecture

A self-reinforcing data flywheel.

The more data Genome accumulates — from public signals, field assessments, and behavioral traces — the better the engine gets. Better signals attract more users. More users generate more data. The flywheel feeds itself.

Data In

Public signals

FRED · EDGAR · BLS · EIA · sector indices

Field assessments

Operators, assessors, consultants — gathering ground truth in exchange for free tools

Behavioral signals

Hiring, content, operator chatter, activity patterns

Expert network

Practitioners contributing domain knowledge, validating signals, executing recommendations

Signal Engine

→ Fingerprint each company

→ Classify archetype + trajectory

→ Score ECL + PFSL overlays

→ Compute conviction signal

→ Connect via knowledge graph

→ Match to experts + actions

Every new data point makes every other signal more accurate. The engine compounds.

Value Out

Investment signals

Conviction scores, archetype transitions, STRONG_BUY/SELL → portfolio alpha

Data insights

Sector health, constraint maps, peer benchmarks → PE, advisors, research firms

Expert matching

Constraint identified → expert matched → execution enabled → transaction fee

Consumer intelligence (Civy)

Simplified insights for retail audience; builds following that feeds more data back in

Agent API

Autonomous agents subscribe to signals, trigger actions — no human in the loop required

Participants

Different participants enter the ecosystem at different points. Some are data contributors. Some are signal consumers. Some are both. The expert network spans all of them.

Data Contributors

Operators & Assessors

Free tools in exchange for ground truth.

Ops Maturity assessments, factory walkthroughs, supplier audits — field observers use free tooling and in return generate the constraint data that makes the engine more accurate. They don't know they're building the dataset.

Signal Consumers

Investors & PE Firms

Behavioral signal before the management call.

Fingerprint 50 targets overnight. Know the binding constraint, archetype, and trajectory before the first conversation. Replace 12-week consultant engagements with a report generated in seconds.

Execution Layer

Expert Network

Constraint identified → expert matched → problem solved.

Every constraint Genome surfaces is a match opportunity. Supply fragility needs a procurement specialist. Labor constraint needs a workforce consultant. Physical constraint needs an equipment dealer. Genome routes the problem to the right expert and takes a fee on execution.

Consumer Layer

Civy / Retail Audience

Industrial intelligence for people who don't know they need it.

Simplified constraint insights, sector health scores, and company behavior patterns — packaged for a retail audience via Civy. Builds a following on TikTok and Instagram. The following generates behavioral signal data that flows back into the engine.

Autonomous Consumers

AI Agents

The end state: no human required.

Agents subscribe to conviction signals via API, execute trades, trigger alerts, route referrals — without waiting for a human to act. The signal pipeline runs continuously; agents are the always-on consumers that eliminate latency between signal and action.

Operating Use

COOs & Plant Leaders

Real-time constraint detection for active operators.

Genome quantifies where constraint energy is concentrated — supply, demand, flow, or cost. Prioritize action before the constraint compounds. Know when trajectory is inflecting before it shows in the numbers.

Products

PE Firms · M&A Advisors

Genome Studio

Diagnostic and benchmarking platform. Fingerprint any industrial company from public signals. Compare against cohort. Identify binding constraint and trajectory. Generate insight-led outreach.

Live (Early)

COOs · Operations Leaders

Operating Command

Decision support for active operators. Detect binding constraint in real time. Simulate intervention outcomes. Recommend prioritized actions with expected impact and time-to-effect.

Planned

Fingerprint Dimensions

Six dynamic properties are computed from the input/output signal pair for each entity. Together they constitute the behavioral fingerprint.

Latency

Response lag

Cross-correlation lag between input and output series. How many periods does the system take to respond to a shock?

Gain

Amplification ratio

Ratio of output % change to input % change. Gain < 1 = constrained response; Gain > 1.3 = fragile/amplifying; Near zero = decoupled.

Damping

Shock decay rate

Computed from detrended residual autocorrelation. High damping = shocks decay quickly (resilient). Low damping = persistence or cycling.

Oscillation

Dominant cycle period

FFT-derived dominant frequency in the output. Present in oscillatory systems (inventory cycles, seasonal patterns). Capped at Nyquist to suppress trend artifacts.

Saturation

Throughput ceiling

Slope-change detection in output as inputs rise. Identifies where output flattens — physical capacity, labor, or regulatory caps.

Vol. Trans.

Volatility transmission

Ratio of output std dev to input std dev (in % change terms). > 1 = amplifies shocks. < 1 = absorbs them. Key fragility indicator.

System Archetypes

Fingerprints are classified into one of seven archetypes using a priority-ordered rule set. Each archetype maps to a distinct management posture and commercial implication.

Constrained

Throughput held back by an internal bottleneck. Responds to inputs but at a reduced rate. Classic constraint signature.

Gain 0.15–0.75 · Low damping · High confidence

Oscillatory

Output cycles with a measurable period. Inventory overcorrection, demand seasonality, or planning lag.

Detectable FFT frequency · Damping < 0.55 · Period < Nyquist

Fragile

Amplifies external shocks. Small input disruptions produce large output swings. High operational risk.

Vol. transmission > 1.3 · Gain > 1.2

Resilient

Absorbs shocks without large output swings. Well-buffered with strong mean-reversion.

Damping > 0.6 · Vol. transmission < 1.1

Saturated

Output plateaus as inputs rise — hitting a physical, labor, or capital ceiling.

Saturation threshold detected · Gain > 0 · High saturation confidence

Decoupled

Output doesn't track macro inputs. Driven by internal dynamics, contract backlog, or signals not yet in the model.

|Gain| < 0.15 regardless of confidence

Constraint Taxonomy (D1–X3)

20-node canonical taxonomy for constraint classification. Each dimension in the fingerprint maps to one or more probable constraint nodes, which drive the recommended actions.

Demand (D1–D3)

End market · Channel mix · Demand stability

Supply (S1–S4)

Raw material · Capacity · Supplier concentration · Workforce

Cost (C1–C3)

Input cost · Conversion cost · Margin realization

Flow (F1–F4)

Inventory · Order velocity · Lead time · Logistics

Risk (R1–R4)

Supply fragility · Regulatory · Financial · Geopolitical

Physical (X1–X3)

Physical constraint · Regulatory cap · Capital constraint

Signal Sources

Three source bundles currently in the model. Each covers a distinct dimension of the fingerprint. FRED is the primary macro layer; SEC EDGAR adds entity-specific financial output; EIA/Baker Hughes adds energy market exposure.

Source	Status	What it provides	Industries / Use Cases
FRED St. Louis Fed public CSV API	Live	Production & Capacity INDPRO · TCU (Total Capacity Utilization) Demand Durable Goods Orders · Retail Sales Cost & Input WTI Crude · PPI (by industry) Flow Inventory-to-Shipments Ratio · Wholesale Trade Labor Manufacturing Employment (MANEMP)	All 16 industry bundles. Primary macro input layer for fingerprinting. No API key required.
SEC EDGAR EDGAR XBRL financial API	Planned	Entity-Specific Financials (quarterly) Revenue (Revenues / SalesRevenueNet) COGS + Gross Margin Inventory Level Capex Operating Income	All publicly-traded industrials. Enables company-level fingerprinting vs. sector-level proxy only. Required for COO and PE diligence use cases.
EIA / Baker Hughes EIA open data · BH weekly rig report	Planned	Energy Market Signals WTI Crude (weekly spot) Henry Hub Natural Gas Brent Crude Baker Hughes North America Rig Count	Agriculture equipment (DE), heavy equipment (CAT), chemicals, oilfield services. Energy cost exposure as constraint input.

IP Boundary

Genome IP (this platform)

System ID algorithms (6 fingerprint dimensions)
Fingerprint schema + archetype classification rules
Peer benchmarking database and methodology
D1–X3 constraint taxonomy
Operating Command action engine
Commercial outreach engine (targeting → outreach → learning)

Freely portable / excluded

FRED signal pipelines (public data infrastructure)
LNS IPI framework, sprint structure — LNS Research IP
LNS gold standards, CDI chat interface
General FastAPI / Next.js patterns