Slack, the popular enterprise communications platform, has unveiled new AI-powered search and summarization features aimed at improving access to institutional knowledge. The platform has long served as a repository for corporate information, but finding and extracting relevant information has been a challenge. With the introduction of these new features, Slack aims to make it easier for users to access and understand the wealth of knowledge stored within the platform.
Noah Weiss, the chief product officer at Slack, explained that the platform naturally accumulates a vast amount of informal and unstructured corporate information. However, surfacing this hidden trove of knowledge has been a complex task. Weiss highlighted the potential of generative AI capabilities to extract meaning and intelligence from the vast amount of data on the platform.
Last year, Slack announced its incorporation of generative AI into the platform with the introduction of SlackGPT, a customized AI model designed specifically for content on Slack. The recent announcement builds on this foundation by introducing more specific applications of generative AI. One such application is the ability to summarize channel content, allowing employees to catch up on conversations or quickly grasp the main points without having to read through lengthy threads.
Weiss emphasized that transparency and trust were key considerations in the design of this feature. Users can request a summary, and Slack’s AI model generates a concise overview of the topics discussed, complete with references to demonstrate how the model arrived at each part of the summary. This transparency empowers users to delve deeper into any area of interest and learn more if desired.
In addition to summarization, Slack now enables users to ask questions in a natural manner, leveraging Slack content instead of generalized internet information. For example, users can inquire about a specific project, and Slack’s AI will provide an answer along with sourcing information to establish credibility. Each answer undergoes a quality check, allowing users to rate its accuracy and providing valuable feedback for model improvement.
While Slack did not disclose specific details about the underlying model, Weiss mentioned that it comprises a combination of large language models. The team invested significant effort in fine-tuning the models to align with the data available on Slack and focused on prompt engineering to enhance performance.
It’s important to note that Slack AI with search and summarization is an add-on product for enterprise plans, entailing additional costs beyond the standard license fees. The feature is currently available in the United States and the United Kingdom, exclusively in English. However, Slack plans to expand language support in the near future.
With these new AI-powered features, Slack aims to revolutionize the way users access and utilize institutional knowledge. By leveraging generative AI capabilities, the platform enables employees to quickly navigate through vast amounts of information, saving time and improving productivity. The transparency and trust built into the system further enhance user confidence in the accuracy and reliability of the generated summaries and answers. As Slack continues to refine and expand its AI capabilities, it is poised to become an indispensable tool for organizations seeking to harness their institutional knowledge effectively.