Rakuten Advertising operates the world’s largest affiliate marketplace, yet the data workflow that underpins pricing and promotion decisions is still dominated by manual scraping and fragmented spreadsheets. We introduce a fully automated pipeline that begins with LLM-driven extraction, channels each crawl through the Model Context Protocol (MCP) for tool governance, elevates the extractor to a first-class agent, and finally distributes structured results across an A2A mesh. The front end of the pipeline couples Mendable AI’s FireCrawl with a lightweight GPT-4 nano checkpoint to translate raw HTML into a validated JSON schema stored in NoSQL. Downstream, autonomous analytics agents consume the same A2A stream to generate trend analyses and anomaly scores, which surface in a modular dashboard whose tiles—Rate Monitor, Competitor Benchmark, Anomaly Alerts, ROI Explorer—can be hot-swapped by registering new agent cards. The architecture cuts data latency from days to minutes, reduces manual collection effort by roughly 90 %, and gives campaign managers real-time, plug-and-play insight into cashback optimisation.
Watch the team present this project at 01:11:07 in the session recording here.
Faculty Advisor
Anil Chaturvedi has over 35 years of professional experience at companies such as AT&T Bell Labs, Kraft Foods, Capital One, and Accenture. He has provided consulting services to Bank of America, Fannie Mae, Johnson & Johnson, and Procter & Gamble. His general research interests include enhancing business value using data science. He has patented and published advanced algorithms for predictive modeling, market segmentation, new product development, product positioning, customer loyalty, consumer promotion mix optimization, and brand strategy. He earned his PhD from Rutgers University and an MBA from IIM Ahmedabad, India.