Orientation

What you're looking at

Montclare AI250 is a live analytical engine, not a static list. This page explains the methodology, the data sources, and how every Bluescore is derived — so every number you see can be traced to its source.

Four things you can do here

One platform, four use cases

Rank

Find the highest-momentum AI companies right now, scored across 6 transparent dimensions.

Compare

Side-by-side Bluescore breakdown across up to four companies in the AI250.

Watch

Track companies you care about and see when their score moves.

Research

Dig into signal provenance and the methodology behind any score.

The engine

8 feature agents, one shared data model

Eight specialized agents record signals on a schedule matched to how fast each one changes.

Agent 1

Funding Intelligence

Recency, size, and stage of capital raises. Last valuation, funding rounds, days since last raise.

Agent 2

Hiring Momentum

Headcount growth, open roles, engineering ratio, layoff events in the last 90 days.

Agent 3

Technical Velocity

arXiv papers, GitHub stars and commits, patents filed, open-source model releases.

Agent 4

News & Sentiment

Media mention volume, sentiment score, velocity trend, tier-1 press mentions, negative news flags.

Agent 5

Partnership Signals

Enterprise partnerships, API integrations, government contracts, distribution partner tier.

Agent 6

Revenue Indicators

ARR signals, revenue growth trajectory, paying customer count, pricing and enterprise tier existence.

Agent 7

Leadership Stability

C-suite hires and departures, CEO tenure, founder-led flag, board expansion, CFO hire signal.

Agent 8

Competitive Position

Category rank, web traffic trend, G2 review count and rating, competitor funding gap, market share signal.

The math

5 steps from raw signal to Bluescore

Every step is documented on the methodology page and applied the same way to every company.

1

Collect

Eight research agents and our data feeds record each company's signals on a schedule matched to how fast they change, from news daily to funding weekly.

2

Coalesce & Smooth

Each signal takes its latest known value, smoothed with a trimmed mean over about seven readings, so a single noisy day can't move a rank.

3

Attribute

Whole-company figures such as revenue, patents and papers are attributed to the AI share: AI Native 100%, Applied AI 60%, AI Adopter 25%, Big Tech 20%.

4

Normalize

Every signal is compared across all 250 companies and mapped to 0–10, then rolled up into 6 weighted dimensions.

5

Weight & Rank

The weighted average of the 6 dimensions is the Bluescore (0–100). Companies are ranked 1–250, every night.

The data

8 signal categories, mostly public

The model is built on public data — every input can be independently verified. Proprietary enrichment is optional uplift, not load-bearing.

01

Funding & Valuation

Crunchbase signals, public filings — who raised what, when, and at what valuation.

02

Hiring & Headcount

LinkedIn public signals, job board data — who's growing and who's contracting.

03

Technical Output

arXiv, GitHub, USPTO — papers, commits, patents, model releases.

04

News & Media

Google News, tier-1 press — volume, sentiment, velocity of coverage.

05

Partnerships

Public announcements, SEC filings, press releases — enterprise and government relationships.

06

Revenue Signals

Pricing pages, customer references, public ARR disclosures, stock price movement.

07

Leadership

LinkedIn, public bios, press releases — C-suite stability and board composition.

08

Competitive Position

G2, SimilarWeb signals, category rankings — relative market standing.

Why trust it

Built to be verified

Every score is kept

Each day's Bluescore and rank are saved, so any past score can be looked up and compared.

No pay-to-play

No company can pay to influence or improve their score. Rankings are editorially independent.

Public & auditable

The core methodology is built on public data anyone can verify. No proprietary black box.

Assumptions on record

Every modeling assumption — weights, lookback windows, smoothing — is documented on the Transparency page.

Feedback