CGA™ Methodology

How CGA™ Works

Our data sources, AI pipeline, and the academic arguments behind every ProIntel score. Scoring logic is proprietary — but the signals and structure are transparent.

19Data Sources
33Sectors
670+Signals Collected
330Problems Scored
47Critical Opportunities
01 — Scoring Engine

Problem-Market Fit Score

PMFS is ProIntel's proprietary scoring model, grounded in Tony Ulwick's Outcome-Driven Innovation (ODI) framework and adapted for AI-driven industrial discovery at scale.

Two-Dimensional Gap Analysis
Problem Importance
How critical, widespread, and urgent is this problem? Derived from signal volume, financial impact mentions, urgency language, and regulatory/safety consequences across 9 data sources.
Solution Satisfaction
How well do existing solutions address the problem today? Measured through complaint frequency, active workaround discussions, product review data, and resolution patterns.
The final PMFS score combines both dimensions through a proprietary weighting model that penalizes over-served markets and amplifies genuine gaps. Problems are classified into tiers — CRITICAL, HIGH, and MEDIUM — based on where they fall in the opportunity landscape.
📐
Grounded in ODI
Ulwick's Outcome-Driven Innovation has a 4× higher success rate than traditional market research. It separates "what people want" from "what they actually need" by measuring outcome gaps, not stated preferences.
🎯
Gap-First Ranking
Problems with high importance but low satisfaction score highest — they represent the largest untapped markets. Problems that are already well-solved are deprioritized regardless of how frequently they appear.
🔍
AI-Calibrated
Each score is AI-assigned based on the actual signal content — not keyword counts or heuristic rules. Claude evaluates context, severity, and resolution status before assigning either dimension.
02 — AI Pipeline

3-Step Hybrid Analysis

We use two Claude models in sequence. Haiku handles bulk work at low cost; Sonnet generates investment-grade reports only for the top 20.

1
Signal Collection
No AI
9 scanners collect raw signals from Reddit, GitHub, YouTube, HN, Stack Overflow, G2, News, Trends, Twitter/Facebook. Output: ~670 signals.
2
Problem Clustering
Claude Haiku
For each sector, Haiku clusters 10 specific, distinct, unsolved problems from raw signals. Output: 330 structured problems.
3
Opportunity Scoring
Claude Haiku
Each problem is evaluated on both scoring dimensions. The proprietary PMFS model converts these into a final opportunity score and tier classification.
4
Investment Reports
Claude Sonnet
Top 20 problems get full 3-paragraph investment reports: Market Pain, Solution Gap, Opportunity. ~700 tokens each.
03 — Data Sources

19 Signal Scanners

Each scanner targets a different layer of discourse — technical complaints, regulatory filings, investment flows, product reviews, academic research. Together they form the most comprehensive industrial signal network available.

Active Scanners
R
Reddit
Community
Targets technical and professional communities per sector. Captures real-world frustrations, workarounds, and unsolved workflow problems from practitioners.
▶
YouTube
Video
Searches for problem-framing content — "how to fix", "biggest problem in", "why X fails". Captures expert discourse and recurring industry pain shared with large audiences.
G
Google Trends
Trend
Captures rising search interest in problem keywords over the past 12 months. Identifies emerging pain before it surfaces in technical communities.
HN
Hacker News
Tech
Filters for high-signal posts from operators and engineers. HN is a consistent leading indicator for enterprise software pain and infrastructure failure patterns.
SO
Stack Overflow
Dev
Targets unanswered questions with high view counts — a direct proxy for unsolved technical problems. Filters by industry-relevant tags and recency.
GH
GitHub Issues
Dev
Targets open issues with high reaction counts. Captures real engineering failures affecting large developer populations — problems not resolved by existing tooling.
N
Google News
News
Scans the last 90 days of industry news. Captures regulatory actions, product recalls, enforcement letters, supply chain failures, and market-level crises.
G2
G2 / Capterra
Reviews
Mines software review platforms for complaint patterns. A high volume of recurring complaints against existing solutions signals a gap waiting to be filled.
𝕏
Twitter / Facebook
Social
Captures real-time industry complaints, product failure discussions, and emerging frustrations from practitioners and decision-makers across sectors.
In Development — Expanding Coverage
in
LinkedIn
Professional
Professional pain point posts, industry frustrations from practitioners, and job-framing content from operators and executives across 33 sectors.
P
Patent Databases
IP
USPTO and EPO filings reveal problems actively being patented around — a strong signal that the underlying challenge remains commercially unsolved.
A
Amazon Reviews
Reviews
Product-level complaint mining across B2C and industrial categories. Recurring 1–3 star patterns at scale indicate systematic market gaps.
Gd
Glassdoor
Workforce
Employee reviews surface internal operational pain — broken workflows, tooling failures, and process gaps that companies struggle to solve at scale.
CB
Crunchbase / Dealroom
Investment
Funding flow analysis reveals which problem areas attract capital. Gaps where funding is absent despite clear pain signal represent the highest-value white spaces.
Rx
ArXiv / PubMed
Research
Academic papers framing open problems are early-stage signals — researchers identify gaps years before commercial products appear. Leading indicator for deep-tech sectors.
AS
App Store Reviews
Mobile
iOS and Android review mining at scale. Captures software UX failures and workflow gaps from real end-users — high-volume, low-friction complaint data.
RF
Regulatory Filings
Compliance
FDA recalls, OSHA citations, EPA violations, and SEC enforcement actions. Compliance failures are leading indicators of persistent, high-stakes unsolved problems.
TP
Trade Publications
Industry
Domain media coverage from Engineering.com, Chemical Week, IEEE Spectrum, and sector-specific journals. Surfaces industry-acknowledged challenges with editorial validation.
JB
Job Boards
Workforce
Unusual hiring patterns reveal operational gaps. When companies repeatedly hire for the same specialist role, it signals a workflow or tooling problem they can't yet automate or solve.
04 — Efficiency

Smart Scanner Routing

Not every scanner is relevant for every sector. Running Reddit for "Oil & Gas" is wasteful — engineers discuss drilling problems on Stack Overflow and GitHub, not Reddit communities. We route intelligently.

Sector Types
industrial Mining, Oil & Gas, Maritime, MRO — GitHub + SO + HN
software Cybersecurity, Enterprise IT — GitHub + SO + Reddit + HN
b2c Hospitality, Retail — Reddit + G2 + Trends + Twitter
b2c_tech FinTech, HR Tech — all scanners active
Why This Matters
Running every scanner against every sector introduces noise — industrial engineers don't discuss drilling failures on consumer social media, and B2C product gaps rarely surface in GitHub issues.

Smart routing ensures each sector is scanned by the sources most likely to carry genuine signal for that domain — reducing noise while increasing relevance and accuracy.
05 — Theoretical Basis

Why This Works

ProIntel is grounded in validated innovation research. We didn't invent the scoring logic — we operationalized it at scale using AI.

"Markets that are underserved are where companies can achieve significant advantage. The key is finding the gap between importance and satisfaction."
Tony Ulwick — Jobs to be Done / Outcome-Driven Innovation (ODI)
"The best startups don't create demand — they find demand that already exists but has no adequate supply. The signal is always in the complaints."
Paul Graham — Y Combinator, "Do Things That Don't Scale"
"Online discourse — Reddit, GitHub, Stack Overflow — is the most unfiltered expression of professional pain. It correlates strongly with market opportunity."
CGA™ Internal Methodology — Signal-to-Opportunity Correlation Study
"Regulatory enforcement actions, product recalls, and compliance failures are leading indicators of underserved markets — not lagging ones."
CGA™ Internal Methodology — Industrial Signal Classification Framework
06 — Coverage

33 Sectors Monitored

Aerospace & Defenseindustrial
AgTech & Farmingb2c
Automotive & EVsb2c
BioTech & Pharmab2c_tech
Building & Constructionindustrial
Chemical Manufacturingindustrial
Cybersecuritysoftware
Ecommerce Logisticsb2c
Energy & Smart Gridindustrial
Energy & Utilitiesindustrial
Enterprise ITsoftware
FinTech & Paymentsb2c_tech
Food & Beverageb2c
Healthcare ITb2c_tech
Hospitality & Tourismb2c
HR & Workforce Mgmtb2c_tech
Industrial Automationindustrial
Logistics & Supply Chainindustrial
Maritime & Shippingindustrial
Mech. & Industrial Eng.industrial
Mining & Metalsindustrial
MROindustrial
Oil & Gasindustrial
Packaging & Containerindustrial
Retail Techb2c
Semiconductorindustrial
Sustainability & Wasteindustrial
Telecommunicationsb2c_tech
Textile & Apparelb2c
Waste Managementindustrial
Water Treatmentindustrial
White Goods & Appliancesb2c
Workforce & HR Techb2c_tech

Ready to Find Your Opportunity?

Enter your profile and let the Matchmaker align you with the highest-fit problems from our database.

✦ Try Matchmaker