Evolving two products into one platform

Assembled started as a workforce management platform for customer support teams. When AI threatened to upend the industry, we needed to move fast and build a new AI agent product. To protect our core business and product from upheaval while we learned, we decided to make it totally separate. Over time though, a real opportunity emerged: we could combine workforce data and insights with AI agent tools, uniquely positioning Assembled to help teams manage people and AI together in a holistic customer support operations platform.

With a loose mandate from the CEO (“we have to put these products together”), I organized a team (product design, content design, UX engineering) to plan, research, and execute the unification, working with a PM and stakeholders from across the company in engineering, marketing, and customer success:

• Navigation overhaul — We used the opportunity to rethink and refresh years of incremental decisions and cruft, designing not just for where we wanted to be today, but also with room for future growth.
• Customer research — We interviewed our customers on how they thought about the two products to validate and iterate on content design and navigation structure.
• Adaptive experiences — We identified duplicative functionality across products and designed unified experiences to adapt based on what each customer had purchased, balancing effort with impact along the way. There was no time to redesign everything from scratch, but every page should at a minimum just make sense. This involved resolving thorny details, particularly around user and company settings, data and reporting, and a complex roles and permissions system.
• In-product marketing — After discovering many customers were unaware of our AI product but excited when they learned, we identified in-product opportunities to surface new products in ways that felt useful rather than intrusive.
• Safe, iterative migration — We made a technical and release plan to migrate two apps into one, all within a unified codebase, without any risky big-bang moments for our existing customers.

The execution happened over multiple quarters. We started with an internal alpha, then rolled out in beta to lower-risk and new customers, progressing eventually to our largest, strategic multi-product customers. This staged approach helped us surface bugs and edge cases early, gave us room to iterate on feedback, and allowed more time to safely deploy complex permission changes that mainly affected the enterprise tier. We avoided serious disruption and generated significant cross-sell interest in both directions: workforce customers curious about AI, and AI customers wanting deeper operational tools. Most importantly, it created a stronger foundation for the product and the business: an improved technical framework, an experience with room for the future, and a product offering that could show the promise of our strategic positioning.