How AI shapes product strategies at Figma, Cloudflare, GitHub and Ramp

SaaS industry leaders from GitHub, Ramp, Cloudflare, and Figma shared their strategies for implementing AI features during the SaaStr Annual’s AI Summit, revealing diverse approaches to product development and integration.

Product implementation strategies: Each company has taken a unique approach to incorporating AI into their existing products, focusing on their core strengths and user needs.

• GitHub’s Copilot has grown to serve 1.8 million users across 77,000 organizations, with success attributed to their developer-centric approach and deep understanding of their user base
• Ramp views AI as an enabler of seamless user experiences rather than the product itself, focusing on making processes “disappear” for their 25,000 customers
• Cloudflare has retrofitted their infrastructure with GPUs across 170+ locations to democratize AI access and support development
• Figma emphasizes augmenting rather than replacing designers, developing AI features through extensive prototyping and testing

Product roadmap adaptation: Companies are adjusting their traditional roadmap approaches to accommodate AI’s rapid evolution and unpredictability.

• GitHub employs “strategic roadmaps” focused on learning objectives rather than fixed deliverables
• Cloudflare utilizes a three-horizon approach spanning immediate needs to long-term research
• Ramp maintains flexibility with quarterly execution plans, embracing iteration and uncertainty
• Figma leverages hackathons to discover and reshape product direction

Team structure and organization: Each company has developed distinct organizational approaches to AI implementation.

• GitHub distributes AI capabilities across all product teams rather than centralizing them
• Figma maintains a mixed structure with specialized ML engineers alongside AI product engineers and designers
• Ramp emphasizes internal tooling with an applied AI team supporting company-wide innovation
• Cloudflare structures their AI efforts in three layers: operational AI, internal tools, and developer products

Quality control and testing: Companies have implemented various methods to ensure AI feature quality and readiness.

• GitHub developed the COFFEE system for compiler offline evaluation
• Figma created custom visual evaluation systems specific to their product needs
• Ramp focuses on rapid iteration and accepts a “winning percentage” approach to feature deployment
• All companies emphasize the importance of extensive testing before releasing AI features

Future outlook and implications: The integration of AI into SaaS products is driving fundamental changes in how these companies envision their future roles and operations, with many predicting AI will become increasingly invisible while enabling more seamless user experiences and breaking down traditional role boundaries in software development and design.