Operating Manual

IndustrialGraph.ai is an industrial-intelligence operating system for the defense and aerospace industrial base.

IndustrialGraph.ai turns fragmented industrial activity — companies, markets, systems, materials, technologies, and the programs that connect them — into a single, queryable knowledge graph. On top of that graph it runs a reasoning engine that produces grounded analysis, tracks live signals, and surfaces consolidation and capability-gap opportunities.

The platform is organized around a simple loop: ingest signals, structure them into the knowledge graph, reason over that structure to produce insights and opportunities, and let the analyst act through missions, investigations, and evaluations. Nothing is a black box — every inference can be traced back to its evidence.

• Who it serves: corporate development, private-capital investors, operating-company strategists, and consultants working the industrial base.
• What it produces: grounded analyses, benchmarks, opportunity theses, and customer-partner evaluations.
• Core principle: evidence in, traceable reasoning out.

IndustrialGraph.ai is the umbrella brand. The platform and its architecture are named with a small family of sub-brands you will see referenced across the product and its documentation:

Every destination lives under one of five concepts in the left rail: Home, Work, Knowledge Graph, Intelligence, and System.

The persistent left rail groups every destination into five concepts. This is the spine of the product — if you understand these five, you can find anything.

The landing surface: a command bar, quick-start workflow cards, and a live briefing.

Home is mission control. It opens with a command bar ("What would you like to accomplish?") that routes a typed objective straight into Start Analysis, a grid of quick-start workflow cards, and an AI briefing tied to your interests and watchlists.

• Command bar — describe an objective in plain language; it hands off to Start Analysis with your intent prefilled.
• Workflow cards — one click into the most common jobs (start an analysis, explore the graph, qualify a customer-partner, read the research).
• Briefing — the latest signals, watchlists, and recent work so you can pick up where you left off.

Start Analysis, Partner Evaluation, Hypotheses, Missions, Investigations, My Workspace, and Watchlists.

The Work group holds the tools you use to actually produce and organize analysis.

• Start Analysis (/briefs) — Describe an objective in plain language and get a grounded, saveable plan — match, buy-and-build, or vertical integration.
• Partner Evaluation (/darwin-evaluation) — Qualify a potential customer-partner — score a candidate across six weighted dimensions into a single recommendation.
• Hypotheses (/hypotheses) — The ideas that matter right now — strategic hypotheses IIOS forms and scores from live evidence, each with a confidence score and trend.
• Missions (/missions) — Strategic objectives that organize investigations, entities, signals, and recommendations.
• Investigations (/investigations) — Structured deep-dives that assemble evidence around a single thesis.
• My Workspace (/workspace) — Saved companies, subscriptions, and your active working set.
• Watchlists (/watchlists) — Companies, markets, and technologies you are tracking.

Explore Graph is the single entry into six connected entity types: companies, markets, systems, materials, technologies, domains.

Everything the platform knows is organized into one connected graph, reached through a single Explore Graph entry (/knowledge). Six entity types are linked to one another so you can move from a domain down to the exact company, part, or technology at issue.

• Companies — Firms in the base
• Markets — Demand areas
• Systems — Platforms & end-items
• Materials — Physical inputs
• Technologies — Enabling capabilities
• Domains — Capability landscapes

These entities are not siloed lists: a company sits in domains, competes in markets, builds systems, depends on materials, and embodies technologies. Signals and reports attach to entities, and analysis traverses the relationships between them.

Signals, Signal Pipeline, Insights, Research, Opportunities, Industrial Pulse, Supply Chains, and Recommendations.

The Intelligence group is where raw activity becomes structured intelligence and where the platform's analytical outputs live.

• Signals (/signals) — A living industrial-intelligence feed classified to the taxonomy.
• Signal Pipeline (/pipeline) — A PR-style review pipeline — ingest articles, run AI entity extraction, then approve signals into the knowledge graph, market bibles, and opportunity engine.
• Insights (/inference) — Inference as a native object — trace how evidence becomes observations, hypotheses, inferences, decisions, and outcomes through the Reasoning Chain.
• Research (/insights) — Institutional reports, each opening with a Decision Maker Summary.
• Opportunities (/opportunities) — Consolidation, integration, and capability-gap theses.
• Industrial Pulse (/observatory) — The macro view of the Industrial Digital Twin — capability, ecosystems, and inflection points.
• Supply Chains (/supply-chains) — Map the North American supplier and manufacturing base — from raw-material producers to OEM integrators — by capability, country, and tier.
• Recommendations (/recommendations) — AI-surfaced next actions tailored to your missions and watchlists. (coming soon)

The intelligence flow, as drawn on the site-relationships map: Signals → Signal Pipeline (review); Signal Pipeline → Explore Graph (commit); Explore Graph → Insights (reason); Explore Graph → Opportunities (surface); Explore Graph → Industrial Pulse (roll up); Start Analysis → Missions (organize).

A natural-language interface grounded only in the platform's own corpus; it can also explain this manual.

The assistant is a natural-language interface to the platform, available from anywhere in the workspace. Ask about a company, program, technology, market, or recent activity and it answers concisely, grounded only in the platform's own corpus — it will not invent companies, numbers, or sources. Relevant entities are rendered as clickable source cards beneath each answer.

The assistant is also aware of this Operating Manual, so you can ask how the platform works — "how do I qualify a customer-partner?", "what is the Signal Pipeline?", "where do I start an analysis?" — and it will explain the relevant area and point you to it.

Engine, Methodology, About, Operating Manual, and Settings — how the platform works and how an administrator runs it.

The System group is for understanding and operating the platform itself.

• Engine (/workbench) — Administrator console to inspect, calibrate, benchmark, and audit the Industrial Reasoning Engine. Nothing is a black box.
• Methodology (/framework) — The Industrial Evolution Framework — the master taxonomy and how intelligence is assembled.
• About (/about) — What this platform is, who it serves, and how the data is assembled.
• Operating Manual (/manual) — How the platform is organized, what each area does, and how the pieces connect — with an interactive site-relationships map. Printable and downloadable.
• Settings (/settings) — Profile, workspace preferences, and notifications. (coming soon)

For administrators: start at Engine to calibrate and audit reasoning, use the Signal Pipeline to govern what enters the knowledge graph, and consult the site-relationships map in this manual to understand how a change in one area propagates to the rest of the platform.

The site-relationships map shows concepts, pages, entities, and the flow of intelligence in one view.

The interactive site-relationships map on this page lays out the whole platform: the five concepts, the pages under each, the six knowledge-graph entities, and the AI assistant. It also shows the direction intelligence flows — from Signals through the Pipeline into the Knowledge Graph, and out to Insights, Opportunities, and Industrial Pulse.

• Solid links show containment — which pages belong to which concept.
• Directed links show the intelligence flow between areas.
• Dashed links show what the AI assistant is grounded on.
• Click any page node to open the real destination.