CoModel AI
CoModel AI turns company data into decision paths by modeling relationships, trade-offs, and system-wide consequences.
We'll email you a 6-digit code to verify it's you. No password required.
CoModel AI
CoModel AI

The Company Model That Makes LLMs Work for the Enterprise.

CoModel AI gives general-purpose LLMs the company-specific intelligence they need to reason, simulate, decide, and act safely.

Schedule a Call
01

Generic AI does not understand your company.

LLMs know language.
They do not know your operating structure.

They do not understand your KPIs, constraints, trade-offs, business rules, decision paths, or internal logic.

That is why generic AI can answer questions, but still struggle to make enterprise decisions.

02

Enterprise intelligence is trapped.

Your company already has a model.

It lives across ERP systems, spreadsheets, dashboards, SOPs, meetings, finance models, and individual experience.

The problem is that this intelligence is fragmented.

  • It is not computable.
  • It is not available to agents.
  • It is not ready for simulation.
  • It is not usable for safe execution.
03

Every enterprise needs a Company Model.

CoModel AI turns business facts into a company-specific reasoning model.

It maps how your company works, what it optimizes for, what constraints apply, what trade-offs matter, and what actions are allowed.

Your LLM knows the world.
CoModel AI knows your company.

Together, they make enterprise AI usable for real decisions.

04
How it works

How CoModel AI works

1

Ingest Company Facts

CoModel AI starts with the operating facts of the enterprise: metrics, entities, rules, constraints, events, levers, objectives, and relationships.

2

Discover Relationships

It identifies how business variables move together, influence each other, and create downstream effects.

3

Map Constraints and Trade-offs

It captures what is allowed, what is risky, and what the company is trying to optimize.

4

Simulate Decisions

Teams can test scenarios before changing capital, policy, process, inventory, staffing, or operating plans.

5

Recommend Actions

CoModel AI compares possible actions and surfaces the expected impact, risk, confidence, and trade-offs.

6

Power Agents Safely

Agents can use CoModel AI as a reasoning layer before they recommend, trigger, or execute actions.

05
Capabilities

What you can do with CoModel AI

Run agents safely.

Give AI agents a company-specific reasoning layer before they act.

Simulate decisions through chat.

Ask what would happen if a policy, process, target, or constraint changes.

Create decision orchestrators.

Turn repeated enterprise decisions into structured, traceable, governed workflows.

Test policies before execution.

Understand likely outcomes before deploying changes into the business.

Build planning and optimization systems.

Move from dashboards and reports to simulation, recommendation, and action.

06

Cloud when allowed. Air-gapped when required.

CoModel AI can pair with local language models so enterprise reasoning stays inside the customer’s environment.

The company model remains the grounding layer.

The language model can change.
The company intelligence stays yours.

  • Use cloud LLMs when speed and frontier reasoning matter.
  • Use private models when control and data isolation matter.
  • Use air-gapped local models when regulation, security, or sovereignty requires it.
07
Examples

Built for high-stakes decisions.

Governments

Simulate economic policies, public programs, infrastructure decisions, budget trade-offs, and downstream effects before implementation.

Financial Institutions

Model decisions across capital, risk, liquidity, compliance, portfolio exposure, and return.

Manufacturing

Optimize inventory, production, fulfillment, working capital, delivery commitments, and operational trade-offs.

Build your Company Model.

Your company already has intelligence.
CoModel AI makes it computable.

Schedule a Call
© CoModel AI · The company model for enterprise AI.
CoModel AI
My HyperIntelligence Models
Open the Sample Model to see how CoModel AI works, or create a new HyperIntelligence Model with your own company data.
CoModel AI
Query Console
Load a model to run deterministic queries.
JSON Inspector
Run a query to inspect the raw JSON request & InsightResponse. The query layer is deterministic — provenance.llm_used = false.
Recent Runs
Loading…
Archive is a local view filter — there is no server-side run delete/archive endpoint, so archived runs stay on the backend.
Influence Graph
No model loaded.
Build or load a run to render the influence graph.
Arrow = direction of influence Green edge = positive relationship Red edge = negative relationship Thicker edge = stronger relationship Fainter edge = lower confidence Direction, sign and strength are read straight from the model — no business "good/bad" is inferred. Technical encodings (W / W_eff / P / confidence C) are in the Advanced tab.
Chat Not connected
Ask questions about this project in plain language. This is a preview of where conversational analysis will live — it is not connected yet and performs no AI or natural-language processing.
💬
Chat is not connected yet.
When enabled, you'll be able to ask about relationships, scenarios, and plans here. For now, use the deterministic Query console on the left.
Details

Company Mode describes your business with six small CSV files. Each file answers one simple question, and together they let CoModel AI map how your metrics influence one another. You don't need every file to be full to start — but the more complete they are, the more the model can do. This guide is static reference material; nothing here is AI-generated.

The six files

Time Periodsobservation_index.csvtime

An ordered list of your time periods — weeks, months or quarters. Every other file points back here to say when something happened. Give each period a stable id and a sequence number so the order is never ambiguous.

Entitiesentity_master.csvobjects

The objects you measure — the company itself, a product line, a store, a region. Each entity has a stable id and a type, and can point to a parent to build a hierarchy.

Metric Catalogmetric_catalog.csvmetrics

A dictionary of every metric you track — revenue, ad spend, churn, headcount. This file defines what each metric means (name, unit, category). It does not hold the numbers themselves — those live in metric_observations.

Metric Observationsmetric_observations.csvvalues

The actual numbers — one row per metric, per entity, per time period. This is the bulk of your data. Each row references a period (index_id), an entity (entity_id) and a metric (metric_id), and carries the value.

Events & Decisionsevents_decisions.csvwhat happened

Notable things that happened — a price change, a campaign, a policy decision — tied to a period. Optional context that helps explain shifts in your metrics. It can start header-only.

Business Factsbusiness_facts.csvgoals · constraints · levers

What you know or want: approved goals (objectives), constraints (limits you must respect), and levers (the metrics you can actually act on). Planning features only switch on once this file holds approved objectives and at least one lever. It can start header-only.

How the files connect

Think of it as a few simple links — each file references the others by stable IDs:

observation_index= time — the calendar every record is stamped against (index_id / sequence)
entity_master= objects — the things you measure (entity_id)
metric_catalog= metrics — the definitions of what you measure (metric_id)
metric_observations= values — every number, joined to a time + object + metric
events_decisions= what happened — notable changes tied to a time (and optionally a metric)
business_facts= approved goals, constraints & levers — the planning inputs

A row in metric_observations is only meaningful because its index_id matches a period in observation_index, its entity_id matches an object in entity_master, and its metric_id matches a definition in metric_catalog. Get those IDs right and everything lines up; misspell one and that number is orphaned.

A monthly business example

Say you run a retail business and track things monthly:

  • In observation_index you list each month: 2024-01 (sequence 1), 2024-02 (sequence 2), and so on.
  • In entity_master you list the company (ent_co) and maybe two stores (ent_north, ent_south).
  • In metric_catalog you define Revenue (m_rev, USD), Ad Spend (m_ads, USD) and Foot Traffic (m_traffic, visits).
  • In metric_observations you record, for each month and store, the revenue, ad spend and foot traffic — e.g. ent_north earned $152,000 of Revenue in 2024-01.
  • In events_decisions you note that in 2024-03 you raised prices 5%.
  • In business_facts you state your goal — grow Revenue — and that Ad Spend is a lever you can move, capped at $50,000/month.

With those six files, CoModel AI can learn how ad spend and foot traffic move revenue, and later suggest how to adjust the lever to reach your goal.

Create your own files — checklist

  • Start with observation_index — list your time periods in order, each with a stable id and a sequence number.
  • List your entities in entity_master — at least your company; add stores, products or regions if you track them separately.
  • Define every metric once in metric_catalog — name, unit and category, with a stable metric_id.
  • Fill metric_observations with the actual values — one row per metric × entity × period, each pointing to the right ids.
  • (Optional) Add events_decisions for notable changes, and business_facts for your goals, limits and levers.
  • Re-use the same ids everywhere — they are how the files link together.
  • Download a template to start from the right columns, or open a completed example to see how a real one looks.

Key things to get right

  • Use stable IDs. Once a period, entity or metric has an id, don't change or reuse it — every other file refers to it. Renaming an id orphans all the rows that point to it.
  • Values go in metric_observations, not the catalog. metric_catalog only defines a metric; the actual numbers always live in metric_observations.
  • Planning needs business_facts. Monte-Carlo planning and the optimizer only unlock once business_facts contains approved objectives and at least one lever. Without them you can still explore relationships, but planning stays locked.
  • Preserve the sequence. Keep observation_index in true chronological order (use the sequence number) — the model reads time order to learn what leads what.

Templates & examples

Don't start from a blank page. Download a blank template (correct headers with a few rows) or a completed example for every file, then adapt it to your own data before uploading.