Skip to main content
Accelerator AI FAQs

Learn more about Accelerator AI and how it can help you to streamline your text analysis.

Trish Pencarska avatar
Written by Trish Pencarska
Updated over a month ago

What is Accelerator AI?

Accelerator AI is our in-platform report-generating tool designed to streamline your text analysis.

It produces comprehensive reports that combine critical evidence, identify key themes and metrics, and provide verbatim examples—all without requiring any effort from the user.

How are the reports generated?

Accelerator AI generates reports through a series of steps:

  1. Theme identification:

    • The tool identifies the top 20 topics, lemmas, and phrases, ranking them by statistical significance.

    • It analyzes the co-occurrence of these topics and lemmas to understand their relationships.

    • The most representative verbatims (direct quotes) are selected based on this analysis.

    • The report then generates between two and four key themes.

  2. Theme reports:

    • Each identified theme is presented with a detailed comparison against other datasets.

    • Metrics for each theme are provided, including relative difference, verbatim frequency, and percentage.

    • The most representative verbatims from both sides of a comparison are included, or from a single dataset if no comparison is made.

  3. Overall comparison/Analysis report:

    • This section offers a condensed version of the theme reports, highlighting the connecting threads between the themes.

    • It focuses on how these themes are used across both datasets.

What are the sections of the report?

  1. Overall differences:

    • A high-level overview of the report that summarizes the key findings.

  2. Dataset breakdown:

    • The report splits the dataset into two sides for comparison:

      • Dataset A: Provides a summary and metrics for the two to four themes identified.

      • Dataset B: Provides a similar summary and metrics for its own set of themes.

Data metrics

What are the metrics used in the report?

  • The report uses the following metrics:

    • Verbatim Level: Frequencies and percentages of direct quotes.

    • Word Level: Relative difference (RD) to compare the significance of words across datasets.

Data security

Does Accelerator AI train on my data?

  • No, Accelerator AI does not train on your data.

Will the AI store our data?

  • No. We use 3rd party GenAI models to create Accelerator AI reports - these 3rd parties have been carefully selected so no customer data is stored by them.

How long will Relative Insight store our data?

  • The Accelerator AI data is covered under our standard data retention policies.

Can I export my data?

  • Exports are a feature that will be available later this year.

Commercial information

  • Accelerator AI is available as an additional module within our platform. For pricing and availability, please contact your account team.

Analytics techniques

Click below for an overview of the various analytic techniques used within the Relative Insight platform.

It covers the creation of analysis reports, the application of machine learning and natural language processing (NLP) methods, sentiment analysis, named entity recognition, and comparative statistical analysis.

The document highlights how these techniques are integrated into different workflow stages to generate insightful reports, incorporating metrics and evidence, with a focus on privacy and security.

If you have any more questions or need further assistance, please contact your account team or drop us a question via the Chatbot.

Did this answer your question?