Slack has been working on a set of AI-powered developer tools that are saving 10,000+ hours of developer time yearly.
Generative AI can have the most impact for developers in areas such as resolving escalations, information search, and assisting in developing AI applications.
For resolving escalations, a bot in Slack categorizes requests and uses Large Language Models (LLMs) to classify posts and escalation requests into specific categories.
In addition, LLMs import data from technical documentation relevant to the channel from various sources, relevant posts inside the escalation channel from history using message search API and can be used to search canvases shared inside the channel using canvas search API.
In order to maximize the effectiveness of classification, each channel has custom configurations, from adjusting prompts, response lengths to setting filters to pick documentation sections from knowledge bases and the ability to respond only when documents surpass a threshold value for similarity scores.
To assist developers looking for information on the web, an interactive web application was developed, primarily for advanced customization and agents have access to internal knowledge sources such as technical documentation.
Slack has actively been partnering with engineering teams internally to embed AI in existing internal tools and to develop new applications and to assist in AI development efforts.
While working with 16+ internal teams to create novel AI experiences, challenges and limitations to integration were noticed, such as the technical limitations of the current generation of LLMs that severely limited their use in complex technical projects.
Slack is currently saving 10,000+ engineering hours by the end of this year across all their tooling with generative AI.
Slack recognized early that the key to success in adopting generative AI lies in balancing the technology with user needs, maintaining a strong focus on security and compliance, and fostering a culture of continuous improvement.