Welcome.AI
    Skip to content
    AI Agents

    AI Surpasses Analysts in Quarterly Earnings Forecasting Accuracy

    AI's predictive power is set to revolutionize earnings forecasts, as demonstrated by Samaya's superior accuracy over Wall Street analysts in predicting major companies' quarterly earnings.

    fortune.com•November 19, 2026•2 min read

    Key Facts

    • AI predictions outperformed analysts for 80% of companies, indicating a shift in forecasting reliability.
    • Samaya's accuracy suggests potential vulnerabilities in traditional analyst consensus methods.
    • Analysts struggled with large firms, revealing a competitive advantage for AI in data processing.
    • Earnings misses lead to disproportionate stock drops, highlighting risks in investor expectations.
    • AI's rise may disrupt the analyst-IR relationship, signaling a strategic shift in earnings forecasting.

    Summary

    Summary

    Samaya, a financial AI startup, aimed to predict quarterly earnings more accurately than Wall Street analysts. In a two-week experiment, Samaya's AI forecasts outperformed analyst consensus for 40 out of 50 major companies during the Q3 earnings season. This result highlights the potential for AI to disrupt traditional earnings prediction methods.

    Background

    Samaya is a financial AI startup that raised $43 million from notable investors in 2025. The company operates in the finance industry, focusing on earnings predictions for publicly traded companies. Before deploying its AI, the traditional method relied heavily on analysts' consensus estimates, which often influenced stock prices significantly.

    Challenge

    The primary challenge was to determine whether AI could provide more accurate earnings predictions than the established consensus from Wall Street analysts. This was crucial as accurate earnings forecasts can greatly impact investor decisions and stock prices.

    Solution

    Samaya implemented its AI model to predict earnings for 50 large companies across various sectors, including health care, energy, and IT. The AI-generated estimates for key metrics like earnings per share (EPS) and revenue were provided at least a day before each company's earnings release. These predictions were then compared to the consensus estimates from analysts supplied by Bloomberg.

    Results

    Samaya's AI forecasts were more accurate than the analysts' consensus for 40 out of the 50 companies. Notable successes included predictions for Verizon and UPS, where Samaya's estimates were closer to actual earnings than those of the analysts. The AI's notable miss was with Yum Brands, where it underpredicted EPS compared to the analysts.

    Key Insights

    The experiment indicates that AI can significantly enhance the accuracy of earnings predictions, potentially disrupting the traditional analyst consensus model. It also raises questions about the biases inherent in human predictions, which AI models may not share, though there is a risk of AI internalizing these biases over time.

    Customer Testimonial

    No direct quote is available from the source material.

    Entities Mentioned

    Companies

    Samaya
    Coca-Cola
    Carvana
    Verizon
    UPS
    S&P Global
    PayPal
    Palantir
    Permian Resources
    Yum Brands
    Amazon
    Google
    Zacks
    FactSet
    Bloomberg

    Technologies

    AI

    People

    Maithra Raghu
    Sheraz Mian
    Paul Hribar

    Key Concepts

    Earnings predictions
    AI vs analysts
    Investor behavior
    Consensus estimates
    Impact of earnings misses
    Corporate investor relations
    Whisper numbers
    AI model training

    Definitions

    Earnings season
    A period when publicly traded companies report their quarterly earnings, typically leading to significant market activity.
    Consensus estimate
    An average prediction of a company's performance based on the forecasts of multiple analysts.
    Whisper number
    An unofficial earnings forecast that reflects what investors believe a company will report, often differing from public consensus.
    AI model training
    The process of teaching an AI system to make predictions based on historical data and patterns.
    Regulation Fair Disclosure
    A rule that prohibits companies from selectively disclosing material information to analysts and investors.

    Use Cases

    • →Predicting quarterly earnings
    • →Assisting investors in decision-making
    • →Enhancing accuracy of financial forecasts
    • →Providing alternative insights to traditional analyst reports
    • →Training AI models with diverse data sources

    Frequently Asked Questions

    How does AI improve earnings predictions?

    AI can analyze vast amounts of data quickly and identify patterns that may not be apparent to human analysts, leading to more accurate forecasts.

    What are the implications of AI outperforming analysts?

    If AI consistently provides better predictions, it could disrupt traditional analyst roles and change how investors approach earnings reports.

    What is a consensus estimate?

    A consensus estimate is the average forecast of a company's earnings as predicted by various analysts, serving as a benchmark for performance.

    What challenges do analysts face in forecasting?

    Analysts often face pressure to conform to consensus views, which can lead to conservative estimates and a reluctance to make bold predictions.

    How might AI change the earnings reporting landscape?

    AI could lead to more accurate forecasts and potentially diminish the influence of traditional analyst consensus, altering investor behavior and expectations.

    Where AI Leaders Stay Informed

    The latest AI intelligence, case studies, and research — delivered to your inbox every week.

    Free to read. Unsubscribe anytime.