Reforge's Playbook for High-Velocity AI Shipping
A new breakdown details how Reforge launched five AI products in just nine months with a small team. Their strategy relied on rapid experimentation, a focus on metrics like A/B tests and cohort analysis, and lean teams where PMs take on overlapping design and data work.
Reforge's pivot to AI tools wasn't just about adding features; it was a strategic decision to create an AI-native product suite covering the entire development lifecycle. This multi-product strategy is a deliberate choice to place several bets on core areas of the product development process rather than aiming for a single "moonshot." The four core products in this suite are Reforge Insights, Research, Build, and Launch. Insights aggregates and analyzes fragmented customer feedback from various sources, while Research uses an AI interviewer to help capture new user insights at scale through voice-based conversations. Reforge Build is an AI-powered prototyping tool designed for product teams to iterate on their existing products, not just build from scratch. Rounding out the suite, Reforge Launch provides AI-native feature flagging and configuration, allowing teams to manage the probabilistic nature of AI products in a live environment. This suite of tools is designed to facilitate a new, faster product development cycle: moving from an idea directly to a prototype for instant learning and iteration. This workflow aims to replace lengthy traditional product requirement documents with functional prototypes that make ideas tangible for teams and stakeholders. The company's rapid pace is underscored by the fact that they also killed a fifth product during this nine-month sprint. This highlights a core tenet of their strategy: a willingness to quickly discard experiments that don't show promise, a crucial discipline in high-velocity development. The structure of Reforge's AI suite points to an evolution in the Product Manager role itself, where PMs are more deeply embedded in the design and data analysis process. By using tools like AI-assisted prototyping and automated feedback analysis, PMs can directly engage in tasks traditionally siloed in design and data teams, accelerating the entire workflow.
