How we used smart RAG and LLMs to resolve merchant matching problems at Ramp in under 10 seconds.

Ramp Engineering's blog offers insights, tutorials, and resources for software engineers and technology enthusiasts. Readers can learn about software development methodologies, engineering culture, and industry trends. With articles, interviews, and technical analysis, Ramp Engineering provides  guidance and expertise for navigating the complexities of the tech industry.

Ramp Engineering

Ramp built an AI agent using LLMs, RAG, and embeddings to automatically fix incorrect merchant classifications for credit card transactions. The system resolves customer requests in under 10 seconds instead of hours, handling nearly 100% of requests compared to the previous 1.5-3% manual resolution rate. The agent uses multimodal context including transaction data, receipt images, and user memos, with careful guardrails to prevent hallucinations and inappropriate actions.

How Ramp Fixes Merchant Matches with AI