Skai’s Celeste AI: Revolutionizing Ad Data Analysis with Amazon Bedrock

Skai, an omnichannel advertising platform, has significantly enhanced its customer insights using Amazon Bedrock Agents to power Celeste, a generative AI agent. Previously, Skai customers faced challenges with time-consuming report generation, difficulty summarizing complex data, and a lack of data-driven recommendations. Celeste addresses these issues by enabling natural language queries to analyze data from various sources, synthesize information into concise narratives, and provide actionable recommendations. Built on Amazon Bedrock, Celeste leverages foundation models like Anthropic's Claude to process queries, access Skai's extensive advertising data, and generate insightful responses, including campaign suggestions and performance analyses. This reduces report generation time by 50%, case study creation by 75%, and overall time-to-value. The adoption of Amazon Bedrock simplified Skai's architecture, reduced development time, and ensured data security and compliance. Key features include a natural language interface, automated data aggregation and analysis, and data-driven recommendations. The target audience is brands and agencies using Skai's platform for omnichannel advertising. Technical specifications include the use of Amazon Bedrock Agents, Anthropic's Claude 3.5, and a custom API for data access. Challenges encountered included latency issues and exceeding token limits, which were addressed through workflow optimization and session chunking. The successful implementation of Celeste demonstrates the power of Amazon Bedrock Agents in democratizing data analytics and accelerating time-to-value for businesses. Future development plans include personalized user experiences, custom data ingestion, and enhanced team collaboration tools.

(Source: https://aws.amazon.com/blogs/machine-learning/skai-uses-amazon-bedrock-agents-to-significantly-improve-customer-insights-by-revolutionized-data-access-and-analysis/)