Eventum engineered a custom large language model (LLM)-powered Slackbot capable of delivering context-aware responses directly within Slack. The solution involved:

• Prompt Interpretation: User queries entered into Slack were processed first by a lightweight LLM (OpenAI's GPT-3.5 Turbo) that quickly parsed the user's message, extracting the specific financial news website URL and isolating the precise question or informational need.
• Dynamic Web Scraping: The Slackbot automatically performed on-demand scraping of the targeted financial news websites using Beautiful Soup + Requests. This allowed for real-time retrieval of text content from diverse, dynamically structured websites, ensuring the extracted data was fresh and contextually accurate.
• Tailored LLM Responses: Scraped data, along with the user's refined query, was fed into a more powerful and sophisticated LLM (GPT-4) that synthesized an accurate, context-sensitive, and actionable response. This enabled analysts to gain deep insights without manually parsing extensive and complex financial news content.
• Conversational Interaction: Integrated directly within Slack, the solution provided an intuitive conversational interface, enabling continuous context-dependent dialogue. Analysts could ask follow-up questions and receive instant clarifications and expansions, significantly boosting workflow efficiency and user satisfaction.

• 90% reduction in the time required for analysts to obtain actionable insights from financial news sources.
• Significantly improved user experience through an intuitive, seamless conversational interface within Slack.
• Enhanced analyst productivity by enabling rapid iterative queries and reducing context-switching.

‍