We started Double in 2018 with the ambitious goal to give people two hours of time back every day, by democratizing access and tools to delegate to a human assistant. Initial research revealed two barriers: finding a great assistant, and building the relationship. We raised a $2M seed round, then later a $8.5M series A to solve this problem.
As Head of Product, my role was to evaluate and prioritize product ideas, then ensure that we were iterating and getting insights fast enough to reach a product market fit. I also led the design side and worked with another designer on a daily basis.
After talking to dozens of executives and assistants, it was clear that not all executives were ready to use a new app, while assistants needed the guidance and access to their executives. We knew that the most streamlined experience would require both assistant and executive to use our tools.
So we built an iOS app for executives, a web dashboard for assistants, and provided fallbacks for key features like ratings over email.
To make the app worth it for executives, delegation needed to be faster than it would be over text or call, while providing more overall value than the existing messaging apps.
We focused the app on voice input, with a live transcript on the assistant side. Each task had a conversation, and executives were only notified of important questions. We also asked them to rate how their week went. This flow required a lot of iterations to find the right tone of voice, balance the need for feedback, and the additional work required by executives.
Most assistants had between 3 and 5 part-time clients, so the need to streamline all their tools was high. Since they primarily worked from their desktop, we built a web app that allowed them to quickly switch between profiles, while notifying them of urgent messages.
Once we had resolved our biggest feature gaps, 50% of executives and 95% of assistants used the apps regularly. While the apps were loved by those who used it, it wasn't enough.
In our interviews with assistants, we understood that some of them were struggling with the expertise required by some tasks. They needed more guidance, and more coaching.
As we scaled the company to onboard hundreds of clients, key operational challenges also emerged: vetting, transitioning assistants, and matching clients were highly time consuming.
With these learnings, in 2021 we decided to re-focus our product energy on supporting operations. To scale the process of vetting, matching, onboarding, and coaching assistants, we invested massively in our assistant dashboard.
In our interviews with assistants and our Operations team, assistants told us they wanted more control on the number and profiles of executives they worked with. As a result, we added a matching catalog to their dashboard, to let them see new anonymized profiles of new executives joining the platform.
Another big challenge was transitions: when an assistant stopped working with a client, or when they took take time away, ensuring a smooth continuity of service for the executive was both a selling point, and a highly manual process. To solve this, we partially automated our transition process, and improved our executives profile in the assistant dashboard.
After this product pivot, it became clear to me that the path forward was to become an operations company with support from product. ChatGPT also opened new opportunities to save assistants time, and that is what the team focused on after my departure.
We started the company under the assumption that a general purpose AI agent able to work on a wide variety of administrative tasks wouldn't be a reality for some time. The last two years have definitely challenged this assumption.