back_office_ops · finance · workflow
Ramp resolves merchant classification corrections in under 10 seconds using an AI agent with RAG and guardrails
Ramp's automatic transaction-to-merchant matching is imperfect, and manual remediation by support and engineering teams took hours per request. Teams could service only 3% of reports in 2023 and 1.5% in 2024, a share that shrinks further as transaction volume grows.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · User submits reclassification request
A Ramp user requests a merchant classification fix by providing a new merchant name, website, and category.
Tools used
LLMembeddingsOLAPRAG
Outcome
The AI agent now handles close to 100% of merchant reclassification requests in under 10 seconds at the cost of cents per request, and improves nearly 99% of transaction classifications according to LLM-judge evaluation.
What failed first
The manual remediation process by customer support, finance, and engineering teams covered only a tiny and declining fraction of requests — 3% in 2023 and 1.5% in 2024 — and cost Ramp hundreds of dollars per resolution.
Results
Time savedless than 10 seconds
Volumefewer than 10%
Cost replacedhundreds of dollars
Running sinceseveral months
Grounding & classification
Source type: technical build writeup
41 fields verified against source quotes.
agentic workflowai agentdata extractionragform submissionreceiptbuilder submittedfailure mode describedmetric backednamed customerproduction runtime claimedtools describedworkflow describedfinancial servicessoftwareaccuracy improvementautomation ratecost reductioncustomer satisfactioncycle time reductiontechnical build writeupback office opscustomer supportfinance opsagentic task executionautonomous resolutionextract classify route