Topic Deltas

Identify granular insights on sentiment changes by topic and importance.
Get Started

Identify the nuanced topics, sentiment changes, and narratives that moves marketsDescription Copy: What companies say on their earnings calls matters to markets. Track it all with unparalleled detail, accuracy, breadth, and speed. Understand with granularity exactly which topics matter most - positively or negatively, on each topic over time. Micro-level insights can be scaled up to sift through 1000’s of companies and industries.

Easily find which companies are out or under-performing their peers either in aggregate or merely on the topics that matter today.

Explore More
J.P. Morgan:  "In this report, we explore the information contained in earnings call sentiment using novel, alternative data sets. Specifically, we look at the relationship between [Accrete’s Topic Deltas] sentiment on capital returns and the subsequent dividend futures returns. Our analysis shows that the effects of management sentiment appear to be moderately significant on dividend futures returns in both statistical and economic terms."

J.P. Morgan

What's Possible

Instantaneously visualize sentiment changes by topic and importance on a company-by-company basis. Quantify and visualize shifts from quarter-to-quarter and year-to-year. Construct intelligent micro-to-macro trend views to reveal the complete picture.

Search S&P 500 company earnings call transcripts going back 10 years. Run advanced queries that can understand the contextual meaning of industry-specific language. Return relevant results quickly that unveil valuable insights for better-informed decision-making.

Individual snippets for any earnings call can be color-coded, removed, and re-categorized to preferred categories. Users can annotate the text with comments, and save everything in a secure company repository.

Enables analysts and associates to focus their attention on ideation, hypothesis testing, and analysis while Topic Deltas prepares the call notes. Replicates the notes that an analyst would take on an earnings call by filtering, extracting and organizing all relevant snippets in intuitive categorizations built for investment analysts.

Conduct research in a granular bottom-up way across all financial documents. Boost research workflows with human-like accuracy scoring sentiment and classifying language into financial topics. Identify key themes, surface valuable insights and reveal hidden patterns into company fundamentals.

Conduct a side-by-side comparison of snippets on a specific topic across multiple companies or multiple transcripts for the same company. With Topic Deltas, finding opportunities and validating the supporting information impacting markets is faster and more objective.

years of historical data
topics covered across 6 broad categories
topic classification accuracy
Human-like sentiment accuracy

Unlock Insights With

Real-time Insights

With live audio transcription, there is no need to wait for earnings call transcripts. Immediately begin your analysis while earnings calls are in progress. Get ahead of the competition and be the first to surface valuable and actionable intelligence.


Using a hybrid of deep learning models comprised of sequential and feed-forward neural networks Topic Deltas can consistently classify sentiment with 96% accuracy.

Industry-Specific Weighting

By adaptively understanding industry context, the machine accurately computes sentiment with a high level of semantic accuracy. It can understand granular concepts and weigh the importance of topics relative to their value within each industry.

Condense Analytical Workflows

Quickly digest and distill thousands of financial documents. Compile hundreds of hours of tedious analysis into minutes with super-human accuracy.

Graphically Transform Text

Topic Deltas assigns a numerical value to the semantic details of language, considering both context and nuance. Users can manipulate graphical representations to find previously indiscernible patterns.

Customizable Technology

Re-train and extend the system to understand a domain of expertise relevant to your business. Financial institutions are using Accrete’s unique abilities to add value within specific company verticals such as ESG, fixed income, internal analyst notes, and more.

Expert Categorization

Topic Deltas replicates how fundamental analysts would build a model by implementing a classification schema that utilizes hierarchical, industry-agnostic categorizations. This schema was developed by industry experts to maximize algorithmic performance and sync up with the method in which discussions are ingested into financial models.


Accrete technology has the unique ability to adapt dynamically. The system is continuously learning and improving as it incorporates new information and contexts, resulting in  a richer understanding and delivering differentiated outputs that yield predictive value.

Explore More Topic Deltas Use Cases

Alpha Generation
Investor Relations
Algorithmic ETF
Portfolio Optimization
Risk Management
Competitive Intelligence

Curious to Learn More?

Get in Touch