How to summarize PowerPoint slides on Google Drive using GenAI

Crafting concise summaries from a lengthy presentation is often a time-consuming task. Consider scenarios like a business needing an executive overview of a strategy meeting, a university professor seeking condensed insights from student presentations, or an organization preparing summarized reports for stakeholders. Manually reviewing slides and distilling key points in such cases can be both tedious and time-consuming.

KNIME Analytics Platform is a data science tool where you use visual workflows to build your applications. It’s open source and free to use. By combining KNIME with GenAI, this task becomes both simple and efficient. You can streamline the process, generating summaries and delivering them as a professional PDF report, all in a few steps.

This blog series on Summarize with GenAI showcases a collection of KNIME workflows designed to access data from various sources (e.g., Box, Zendesk, Jira, Google Drive, etc.) and deliver concise, actionable summaries.

Here is a 1-minute video that gives you a quick overview of the workflow. You can download the example workflow here, to follow along as we go through the tutorial.

Let's get started.

Our goal is to address the challenges of manually reviewing and condensing lengthy slide decks into concise, informative insights through an automated solution, saving both time and effort.

We can do this in three steps:

1. Access PowerPoint slides by connecting to Google Drive
2. Summarize slides using LLMs
3. Compile summaries into a report and upload it to Google Drive

We begin by accessing the PowerPoint slide decks stored on Google Drive. To do this, we use the Google Authenticator node to authenticate to our Google account, selecting the appropriate scope and permissions such as, read and write access to Google Drive.

Next, we use the Google Drive Connector node to specify the working directory where the files are located. In our example, the slide decks are named Hypothesis Testing.pptx and Game Theory.pptx.

To extract the content of these presentations, we first move the files from the original location to a new destination using the Transfer Files node. This node automatically adjusts its input port based on the source file system we connect to. In this example, we move the files from Google Drive to the workflow data area for local processing.

Then, the Tika Parser URL Input is used to read in the textual content from the .pptx files. Unlike the standard Tika Parser node, this node allows us to select an input column containing the file path, offering more flexibility and opportunities for automation.

Once the slide content is extracted, we use the KNIME AI extension to select the most suitable LLM for the task, balancing costs and performance. For example, OpenAI’s GPT-4o is one option. Other alternatives are also possible, including open-source, local models.

To establish the connection: 

1. Input the API key in the Credentials Configuration node
2. Authenticate to the service using the OpenAI Authenticator node 
3. Connect to the GPT-4o model using the OpenAI Chat Model Connector node

After establishing the connection to the model, we proceed to engineer a prompt using the Expression node.

A simple but effective prompt is crafted based on prompt engineering best practices. The goal is to guide the model towards generating  summaries that are clear, well-structured, and relevant to the user’s needs. Here’s the prompt we used:

join("\n\n", "Summarize the following text as a concise paragraph. Additionally, extract the key five topics as a numbered list: ", $["Content"])

The LLM prompter node then sends the query to  the LLM, generating the slide summaries.

To compile the generated summaries into a static PDF report, we use the KNIME Reporting extension and the component’s composite view.

In the “Slides Summaries” component, we display the LLM responses and design the layout to combine the summarized slide decks with an introductory heading, creating a more engaging and structured report.

The component receives input also from the Report Template Creator node, which defines the page size and orientation of the final report, and outputs to the Report PDF Writer node which saves the report in the workflow data area.

Finally, the report is uploaded to the same Google Drive folder where the slide decks are stored. This is done by using the Transfer Files node, which uses the established file system connection (Google Drive) as a file destination.

This streamlined process makes it easy to share the report with colleagues or archive it for future use.

The workflow generates a clean, well-structured PDF report that effectively presents slide titles and summaries, streamlining information access and highlighting key points as a numbered list.

In this article from the Summarize with GenAI series, we demonstrate how KNIME and GenAI can automate tedious and time-consuming tasks. By accessing PowerPoint slides stored on Google Drive, using LLMs for summarization, and producing ready-to-use reports, the workflow highlights how visual tools and AI can save time, extract valuable insights and enhance productivity.

You learned how to:

• Connect to Google Drive and access PowerPoint slides
• Use the KNIME AI extension to generate concise summaries
• Compile summaries into a PDF report and upload it to Google Drive