Case Study — August 18, 2025
An Investment Firm Uses GenAI to Find Market-Moving Insights in Earnings Calls
THE CLIENT:
A quantitative investment firm
USERS:
Quantitative analysts
AI is no longer a futuristic tool — it is here, helping investors make more informed, data-driven decisions in real-time. From understanding current market dynamics to forecasting future performance, AI frameworks are showing they are game changers in the world of equity analysis. One area where this transformation is especially evident is in how investment professionals analyze earnings calls.
With the pace of today’s markets, the ability to extract meaningful insights from complex, unstructured data—quickly and at scale—is becoming essential. For firms seeking a competitive edge, traditional models and manual research methods can’t keep up. Instead of spending time combing through entire transcripts, firms are now turning to sophisticated AI tools. These tools adapt to each client’s use case and analyze vast amounts of unstructured data to surface key insights.
The quantitative analysts at this buy-side firm had been tracking the evolution of AI for assessing earnings calls. They realized that new capabilities were quickly taking hold, and they needed to advance their assessments to maintain a competitive edge.
Pain Points
The quantitative analysts at this buy-side firm knew they were at a disadvantage without leveraging the latest capabilities for analyzing earnings calls. They were familiar with S&P Global Market Intelligence's ("Market Intelligence") Machine Readable Transcripts that aggregates data from earnings calls delivered in a machine-readable format for Natural Language Processing (NLP) applications with metadata tagging.
They were excited to learn that the company had acquired ProntoNLP early in 2025 to expand its Generative Artificial Intelligence (GenAI)-powered product portfolio and bolster its textual data analytics capabilities.
They were impressed that ProntoNLP:
- Utilizes proprietary NLP capabilities coupled with large language models (LLMs) for fast, efficient and deep analysis of unstructured financial data at scale.
- Efficiently processes and extracts insights to build investment strategies.
- Offers LLM-driven structured dataset on event detection and contextual sentiment scoring.
The Solution
The analysts wanted to learn more about the capabilities, how the data is delivered and the types of insights they could glean. They reached out to Market Intelligence to learn more, where specialists introduced the team to ProntoNLP, a leading GenAI tool that enables users to derive differentiated insights from unstructured textual data. By integrating signals from ProntoNLP's cutting-edge LLM —trained on Market Intelligence data with human experts-labeled paragraphs—users can access richer context, improve prediction accuracy, and uncover insights that generate uncorrelated alpha.
Pronto uses a purpose-built LLM optimized for finance. Each sentence in a text document is labeled separately and then linked to the surrounding information for context. This approach provides details on the importance of a sentence as well as its sentiment. LLM-driven outputs can easily identify context and assign labels such as product names, sentiment, and event tags—capabilities that traditional NLP models often struggle with.
The specialsits explained that datasets are created that enable users to easily perform market research, confirm investment hypotheses and generate highly precise and valuable NLP-based signals without the need for any coding skills. The data can be accessed via numerous options, including a robust user interface and would enable the team to:
|
Turn Transcripts into Signals |
ProntoNLP Transcript Analytics is a powerful new tool for analyzing Machine Readable Transcripts data. By using NLP and an optimized LLM, ProntoNLP accurately recognizes and scores important phrases, helping users easily identify valuable information and separate it from the noise. With this dataset, users can:
|
|
Extract Intelligence |
ProntoNLP Filings Analytics applies the same AI-driven approach to Machine Readable Filings, helping firms uncover risks, challenges, and strategic developments disclosed in company reports. It can be used to identify and rank companies based on challenges and risks mentioned in their filings and reports. It also supports easy analysis of company reports and filings over time and identify inflection points in company strategy and risk. Key features include:
|
|
Dig deep without the need to code |
Analyzed data is accessible through a user interface, APIs, XpressfeedTM (Market Intelligence's data feed management solution) and cloud-based Snowflake. With just one click, users can view a ranked list of companies by their overall sentiment score or by a specific topic and easily identify companies that have positive or negative statements related to specific topics. They can also follow companies’ sentiment change between periods and see which had the most positive change versus the most negative change.
|
The quantitative analysts saw that the offering would enable them to delve into vast volumes of data, unveiling hidden insights and correlations that were previously inaccessible. They saw it as a way to elevate their investment strategies by giving them access to ready-to-use data. This data identifies, extracts, clusters, and scores phrases—spoken or written—with contextual awareness to separate signals from the noise.
They subscribed to the offering and can now rely on trusted data and an advanced approach that:
- Covers 110 key topics, spanning everything from revenue, margin, profit to new product, customer behavior, competition, headwinds and supply chain.
- Updates every hour in a clean structured format, saving time and effort building pipelines and analytics in-house.
Content Type
Theme