May 22, 7:33 am
Our AI Score tracks over 25 alternative data signals across thousands of publicly traded companies, updated daily. It measures what's actually happening inside a business, from hiring velocity and employee sentiment to social media growth and web traffic, and distills it into a single score from 1 to 100.
This article walks through the methodology behind the score, the data that powers it, and the real-world results it has produced. We believe investors deserve to know how a stock-picking system works before they trust it. So we're showing you everything: the inputs, the process, and the receipts.
The AI Score is a composite built from four distinct dimensions. Each one captures a different angle on a company's trajectory, drawing from data sources that traditional financial analysis doesn't cover. Together, they form a picture that earnings reports alone can't provide.
Audience
Social media follower growth across Instagram, X/Twitter, YouTube, Tiktok, Facebook, etc. A company gaining followers month over month is exhibiting demand-side momentum that often precedes revenue acceleration. A company losing followers is a warning sign, as demand might be cooling.
Customer
Web traffic volume and trends, mobile app downloads across iOS and Android, and social sentiment from Reddit and other online stock forums. When app downloads spike or web traffic breaks above a sustained trend, that's often visible in our data weeks before it shows up in a quarterly earnings call.
Employment
Job postings from major online job boards, headcount trends, and employee review signals: satisfaction scores, CEO ratings, and the percentage of employees who expect business to improve over the next six months.
Fundamental
Year-over-year revenue growth, analyst consensus, financial health scores, insider buying patterns, analyst price targets relative to current price, and recent stock price momentum. These anchor the score in financial reality and prevent the model from overweighting hype signals in stocks with deteriorating financials.
Every signal is scored on a standardized scale. We don't just look at the current value. For each input, we measure three things: the absolute level (is this reading strong or weak?), month-over-month momentum (is it improving or deteriorating?), and the long-term trend (has the direction been consistent over a longer time frame?). This triple lens is important. A company with strong job posting numbers that are declining month over month is in a very different situation than one with moderate numbers that are accelerating.
The individual signal scores are then combined into a weighted composite. Signals with stronger historical predictive value carry more weight in the final calculation. The result is normalized to a 0-to-100 scale.
The score updates daily as new data is gathered and analyzed. A stock can move from Hold to Buy over a few weeks if multiple signals start trending in the same direction, or it can fall from Buy to Sell if the alternative data deteriorates ahead of a public catalyst.
On September 2, 2025, we pulled the stocks with the highest AI Scores in our system: the ones where the data was most uniformly positive across all four pillars. These were the highest-conviction picks, stocks scoring 73 or above out of 100. Seven stocks qualified. Here's how they performed over the following nine months.
| Company | AI Score | Sep 2 Price | May 20 Price | Return | |
|---|---|---|---|---|---|
 |
Eli LillyLLY |
79 | $729.04 | $1,018.87 | +39.8% |
 |
CamecoCCJ |
75 | $86.15 | $104.02 | +20.7% |
 |
Bread FinancialBFH |
74 | $65.74 | $87.69 | +33.4% |
 |
VertivVRT |
73 | $138.18 | $315.67 | +128.4% |
 |
CelesticaCLS |
73 | $248.99 | $346.48 | +39.2% |
 |
OneokOKE |
73 | $69.66 | $92.15 | +32.3% |
 |
GapGAP |
74 | $23.42 | $22.30 | -4.8% |
| Portfolio Average (7 stocks) | 74.4 | +41.3% | |||
| S&P 500 (SPY) | $640.00 | $736.50 | +15.1% | ||
|
+41.3% Avg return, top AI picks |
+26.2% Alpha vs S&P 500 |
6 of 7 Stocks beat the S&P |
86% Win rate |
The seven stocks that scored 73 or above in September weren't randomly distributed across industries. They shared a common profile: strong or accelerating alternative data across multiple pillars simultaneously. That's the critical insight behind the AI Score. A high reading doesn't mean one signal was screaming buy. It means several independent data sources, covering different facets of the business, were all pointing in the same direction at the same time.
Vertiv (VRT) was the standout at +128.4%. In September, the company's job postings were rising, its web traffic was accelerating, and its financial metrics were improving quarter over quarter. The AI infrastructure buildout that would eventually drive the stock from $138 to $315 was already visible in the alternative data months before the earnings beats that Wall Street reacted to.
The chart below illustrates the dynamic. Vertiv's AI Score first crossed into Buy territory (above 60) in mid-July, when the stock was still trading around $128. By August, it broke above 70, and it peaked at 80 in late January 2026 with the stock at $182. As the price surged over the following months, the AI Score moderated, factoring the rapid appreciation into its risk assessment.
AltIndex AI Score Data
Eli Lilly (LLY) scored highest at 79 and returned +39.8%. The company's social media growth was strong, its employee business outlook was elevated, and the fundamental picture was improving as Mounjaro and Zepbound revenues accelerated through the second half of 2025. The AI Score caught the convergence of demand signals and financial execution before LLY broke out of its mid-2025 trough near $623.
Gap (GAP) was the only negative performer, down -4.8%. This result is worth examining because it shows what the model gets wrong. GAP's social signals and employee data were positive in September, but the company's web traffic and revenue trajectory were softer than the other top picks. The score of 74 barely cleared the threshold, and the weaker customer pillar proved more predictive than workforce optimism. Not every high-scoring stock wins, but the portfolio-level performance across all seven picks shows the edge is real and statistically meaningful.
Traditional stock analysis focuses on data that's already public: earnings reports, SEC filings, analyst ratings. By the time those data points are available, the market has already priced them in. Alternative data captures what's happening between earnings reports, in the weeks and months before the next 10-Q filing reveals the trend.
When a company's job postings surge 20% in a single month, that often signals product launches, geographic expansions, or operational pivots that won't show up in reported revenue for two or three quarters. When employee business outlook drops from 70% to 40%, that's an early warning of internal problems that could eventually surface as restructuring charges, executive departures, or earnings misses.
Our AI stock picker synthesizes these signals automatically. Instead of manually checking a company's LinkedIn headcount, Glassdoor reviews, Instagram follower trajectory, and web traffic patterns, the AI Score monitors all of it daily across thousands of stocks and distills the result into a single actionable number. For more examples of how our AI picks have performed across different time periods, see our full AI stock picks tracker, which includes historical picks, current top-scored stocks, and real-time performance data.
The highest-conviction picks from September 2025 returned an average of +41.3% over the following nine months, outperforming the S&P 500 by more than 26 percentage points with an 86% win rate. The system works best at the extremes, where the AI Score identifies broad convergence across multiple independent data signals. That doesn't mean every pick wins. It means the methodology, when conviction is high, has demonstrated a measurable and repeatable edge.
See What the AI Score Is Picking Right Now
Track AI Scores, alternative data signals, and stock picks for 3,000+ stocks.
Get Started for Free →Sources: AltIndex AI Score data, AltIndex alternative data signals (job postings, social media, web traffic, employee sentiment), S&P 500 pricing data. Stock prices are historical closing prices and for informational purposes only.
Disclosure: This article is for informational purposes only and does not constitute investment advice. AltIndex aggregates publicly available alternative data signals. Past signal performance does not guarantee future results. Always do your own due diligence before making investment decisions.
Sign up and get access to a personalized dashboard, deeper insights, AI stock picks, stock alerts, weekly newsletter and much more.
