Productizing AI/ML models for Proptech leader, Tailorbird

Tailorbird's proprietary AI/ML model enables reality capture from existing photos and floor plans. This helps property management groups cost out renovation and interior design projects to the finest detail. Zemoso worked with Tailorbird's engineering and design teams to build a prototype that could generate leads, attract investors, and sharpen product-market fit. We launched V1 in four months. From there, we kept evolving the design and features through a fast feedback loop to drive adoption.

• Phase 1: We ran a Google Ventures design sprint to create high-fidelity, clickable prototypes. We co-created the golden path for key users as part of that. The prototype generated leads and helped with product discovery.

         Leveraging the early customers acquired, Tailorbird:
        1. Iterated the prototype quickly using customer feedback with Zemoso’s product pod
        2. Reduced the timeline to acquire proof of value
        3. Raised seed funding from Moderne Ventures with the prototype

• Phase 2: We worked closely with our stakeholders to help productize the AI/ML model and outline the experience around it on an expedited timeline.
  The Zemoso expertise: We provided them with an agile team that self-organized and evolved based on development and design needs.
• Key highlights of the solution: As the first step to laying out a detailed UX design and the system architecture, we aligned with Tailorbird leadership on platform essentials:
  1. Make the platform intuitive and easy-to-use
  2. Deploy user controls
  3. Enable quick release of new features based on user feedback

Scalable architecture and consistent design: The product uses a microservices architecture to make it more robust, ensuring the system is scalable and that any faults introduced during new feature rollouts are isolated and quickly resolved. Alongside this, we co-created the detailed design for all user flows and helped set up access controls. For the front-end, we opted for a Material-UI design approach to deliver a seamless, adaptable, and uniform experience, applying atomic design principles to make releasing new updates and features more efficient.

Flexible data handling and streamlined deployment: Together with Tailorbird’s internal team, we ensured that the product has superior abilities to store and process flexible document schemas (spreadsheets, images, etc.). A GraphQL interface layer was used to ensure easier testing for new features and faster processing times. DevOps practices were used to automate testing and integration and containerization for easier deployment.

Modular architecture and real-time alerts: An integration with Twilio SendGrid ensured that all internal and external stakeholders got alerts within three seconds of an action being taken. The modular approach to building the entire product had another benefit. As most capabilities offered are built in their own self-contained units, property management companies can pick which services to buy and use, encouraging trial and early adoption of the product.

P.S. The partnership with Tailorbird is ongoing and evolving.

By combining a microservices architecture with a modular, atomic design system and fast, GraphQL-driven testing cycles, Zemoso helped Tailorbird turn a prototype into a product that landed three of the top five multi-family property management companies in the US as customers within six months of launch.