Ramp
Ramp is an American financial technology company specializing in spend management solutions for businesses. Its core products include corporate credit cards, expense reimbursement management, bill payment, and financial automation services, designed to help companies control costs and improve financial efficiency. Ramp primarily serves small and medium-sized enterprise customers, simplifying routine corporate financial workflows through the integration of software and financial services.
Timeline (last 90 days)
Ramp's latest AI Index shows token spending declined 5.2% month-over-month, while token usage is rising.
Ramp's economists noted that the decline in token spending is almost entirely driven by competition between OpenAI and Anthropic, with open-source models still accounting for less than 5% of enterprise spending.
Ramp developed a coding agent called Inspect that runs in Slack, responds within 5 seconds, and can return interactive deployment previews.
Ramp built an AI agent called Glass that connects to Snowflake data and understands user research, product strategy, how product specifications are defined, and the entire codebase.
Ramp built a browser-based QA agent called Testo that launches 100 combinations based on real production data to simulate real user actions, catching 425 bugs in the past 30 days.
Ramp's Inspect has accumulated millions of sessions, with 75% of PRs built by it, and 1,000 PRs submitted by non-engineers in the past month.
Ramp built a customer insights agent that uses ETL and data pipelines to pull data sources from across the company, combined with vector retrieval and clustering to understand products, teams, and features.
A context window of one million tokens represents less than 0.5% of Ramp's entire Gong call records.
85% of questions received by Ramp's PMs are answered in full by AI.
Ramp's Gadget can understand the intent of a question and connect to formal records such as Notion roadmaps, specs, customer calls, Slack, and Linear tickets.
37 more timeline events
All Facts (20)
Ramp团队采用分层审查的人机协作模式，AI作为代码审查的第一道防线，人类审查者聚焦架构合理性、业务逻辑正确性和安全合规等高层次判断
80%UnverifiedAccording to Ramp data, 80% of enterprise revenue for OpenAI and Anthropic comes from just 1% of customers.
60%UnverifiedRamp is an enterprise spend management platform that provides services including corporate credit cards, expense management, and procurement.
60%UnverifiedRamp's latest AI Index shows token spending declined 5.2% month-over-month, while token usage is rising.
50%UnverifiedRamp's economists noted that the decline in token spending is almost entirely driven by competition between OpenAI and Anthropic, with open-source models still accounting for less than 5% of enterprise spending.
50%UnverifiedA context window of one million tokens represents less than 0.5% of Ramp's entire Gong call records.
50%UnverifiedRamp built a customer insights agent that uses ETL and data pipelines to pull data sources from across the company, combined with vector retrieval and clustering to understand products, teams, and features.
50%UnverifiedFrom June 2025 to May 2026, Cursor's share of enterprise AI coding tool spending tracked by Ramp fell from 41% to 26%, and JetBrains data showed workplace adoption rates declining from 18% to 12%.
50%UnverifiedThe conversation content comes from a Ramp × GPT-6 Astra demo video on YouTube and represents a vendor-perspective demonstration.
50%UnverifiedOne of Ramp's core value propositions is simplifying accounting workflows through automation, and computer use is key to achieving this.
50%UnverifiedRamp uses computer use technology to fetch receipts from merchant sites in the background for customers, but requires extensive manual prompting to guide the model's actions on various merchant websites.
50%UnverifiedElevenLabs initially integrated financial automation platforms such as Ramp to address bill payment issues, and subsequently expanded to other financial modules.
50%UnverifiedRamp converts unstructured sales data (call records, emails, notes) into vectors via an embedding model and stores them in a vector database to support semantic retrieval
50%UnverifiedRamp built a GTM orchestration system that can automatically distribute actions to execution channels such as outbound calls, advertising, the official website, and in-app notifications by describing intent
50%UnverifiedRamp built a customer data platform (CDP) in-house that integrates CRM data, product data, enrichment data, web data, and purchase signals
50%UnverifiedRamp chose to build its CDP in-house rather than adopt third-party solutions such as Segment or mParticle due to the complexity of its data model
50%UnverifiedRamp uses a Postgres transactional layer to maintain transactional guarantees and referential integrity between entities, and retains complete metadata
50%UnverifiedRamp uses DBT together with the Snowflake cloud data warehouse for offline batch processing computation, and then flows data back through reverse ETL
50%UnverifiedThe core of Ramp's methodology is to first solve the problems of one team and then replicate horizontally
50%UnverifiedDuring Ramp's data hydration process, the same email address may represent multiple companies, so fuzzy matching is required
50%