Conversational AI Agent for Financial KPI Reporting

Challenge

Spoom, the business intelligence platform developed by EASI, is designed to offer fast, user-friendly access to financial reporting. While the platform includes a comprehensive suite of dashboards and KPI summaries, EASI began exploring how to make insights even more accessible, by allowing users to ask questions in natural language. The goal was to streamline access to key metrics without relying on filters or predefined configurations, enhancing Spoom’s value proposition through a more intuitive, conversational experience.

Nature of collaboration

EASI partnered with Sagacify to co-develop a generative AI-powered chatbot integrated into the Spoom platform. EASI developed a dedicated API endpoint designed specifically for this project, enabling the chatbot to access accurate, up-to-date figures using predefined queries that followed Spoom’s existing data rules. Sagacify designed and implemented the generative AI agent, including the logic to understand user questions, identify the correct KPIs or dimensions, and retrieve the relevant information via the API.

Solution

Sagacify developed a generative AI agent that allows Spoom users to ask objective, KPI-related questions and receive accurate, data-driven answers. The agent accessed business metrics through a set of predefined functions connected to Spoom’s API, ensuring results remained aligned with existing calculation logic.

To guide its reasoning, the agent used the ReAct prompting approach—breaking down each query into steps: interpret the question, identify the correct KPI, and call the appropriate API function. This tool-based setup helped control how the AI interacted with the data, reducing the risk of incorrect or unsupported queries.