Inside Ramp, the $32B Company Where AI Agents Run Everything | Geoff Charles

The traditional roles in product development are shifting dramatically. What was once a human-centric process of writing detailed specifications for engineers is now increasingly handled by AI. This fundamental change is accelerating the pace at which products are conceived and built, blurring the lines between different functions within a company.

At Ramp, AI isn't just a helper; it's a co-creator. Currently, an astounding 50% of Ramp's codebase is generated by artificial intelligence. This represents a significant leap from 30% just last December, with projections indicating it will reach 80% by March. This rapid integration highlights AI's transformative power in accelerating software development cycles.

Product Managers are no longer just writing specs; they are actively building and shipping features themselves. With AI tools, they can create full-stack products, integrating both front-end and back-end functionalities. This direct involvement allows for rapid iteration and validation, transforming PMs into hands-on creators who can quickly bring ideas to life.

The impact of AI at Ramp isn't confined to engineering or product teams. Designers, operators, account managers, and sales professionals are all leveraging AI to enhance their work. The core philosophy is to automate every job, freeing up human talent to focus on higher-level, strategic tasks that require creativity and complex problem-solving.

Ramp, a rapidly growing and AI-native company, uses AI to deeply understand its 50,000+ customers. By processing vast amounts of data from sales calls, CRM notes, in-app surveys, support tickets, and chat logs, AI helps identify key customer pain points and product opportunities. This "Voice of the Customer" (VoTC) approach provides unparalleled insights.

Product managers can query the VoTC Agent in Slack, asking specific questions about customer feedback related to any product area, like "procurement." In just minutes, the AI sifts through 90 days of support tickets and chat logs, extracting key takeaways, identifying common themes, and linking directly to the underlying source documents. This dramatically reduces the time spent on research from days to mere minutes.

Ramp Research is another AI agent that provides instant data analysis. Users can ask complex data questions in natural language, and the bot connects to Ramp's entire database, understanding schemas and generating accurate answers with interpretations. For example, a query about the open rate of automated emails across various templates is answered within two minutes, providing detailed metrics and insights that would take a human analyst days to compile.

Ramp utilizes a system of "AI skills" for various tasks, from problem identification and research to product shaping and specification. These skills, often built by PMs themselves, embed best practices and context, allowing AI agents like Claude to act as thinking partners. The ultimate workflow aims for a "prompt-to-product" cycle, minimizing manual steps and maximizing efficiency.

Inspect is a powerful tool that allows product managers to build user interfaces directly from natural language prompts. Instead of writing detailed specs for engineers, a PM can describe desired metrics and a dashboard layout, and Inspect will create a fully functional product. This bypasses traditional front-end and back-end development, drastically shortening the time from idea to a working product.

After a simple prompt outlining desired metrics, Inspect immediately began working. In just five minutes, it delivered a fully functional Bill Pay dashboard with "Past Due," "Upcoming Bills," and "Total Amount Outstanding" metrics. This isn't a prototype; it's a live product, complete with integrated backend and frontend functionality. The process also automatically creates a pull request for engineering review, streamlining deployment.

With AI automating much of the traditional building and specification tasks, the Product Manager's role is evolving into a more strategic and mentorship-focused position. They need to become exceptional builders themselves, guiding teams on vision, strategy, and understanding complex trade-offs. The emphasis shifts from simply managing processes to deeply understanding how to win in the market and achieve business impact.

To thrive in this AI-driven landscape, individuals and companies must prioritize continuous learning and growth. The focus is on cultivating a "beginner's mindset" and the humility to embrace new technologies. Ramp encourages its employees to get deeply embedded in AI tools, recognizing that personal development in this area is crucial for staying ahead and contributing effectively.

Ramp has democratized access to AI tools across the entire organization. There are no restrictions on who can use advanced AI capabilities like Claude Code or Inspect. By removing constraints around access, tokens, and budgets, Ramp fosters a culture of innovation. This allows everyone, from finance to legal to marketing, to leverage AI to automate tasks and build solutions, pushing the boundaries of what's possible.

To inspire widespread AI adoption, Ramp actively showcases non-builders creating innovative solutions with AI. This includes the finance team automating treasury management, the legal team reviewing contracts, and marketing teams generating website content. These internal success stories demonstrate the practical benefits of AI across various functions and encourage others to explore its potential.

The role of management is undergoing a significant transformation. Instead of focusing on task allocation and oversight, managers must now prioritize coaching and development, helping their teams adapt to new AI-driven workflows. This is a challenging shift, as managers must guide teams through skill set changes, sometimes without having those new skills themselves. The emphasis is on fostering growth and ensuring everyone contributes effectively.

In this dynamic environment, the ultimate measure of success is the ability to build great products faster and drive meaningful metric movement for customers. With AI making code development nearly free, the focus shifts to strategic direction, rapid iteration, and eliminating bottlenecks. This is a challenging but exciting time to be a builder, where vision, taste, and the ability to leverage new technologies are paramount.