OpenTable . AI Assisted Review Responses
Designed OpenTable’s first Gen AI tool that generates personalized responses to diner reviews, saving restaurants time while still providing a personal touch.
Team
1 designer
1 product manager
1 product marketing manager
6 data scientists
10+ engineers
Impact
54% of replies leveraged our genAI (24% above goal)
OpenTable is the first amongst it’s competitors to release a genAI feature
My role
I synthesized past research and worked with the PM to lead workshops establishing the product strategy and defining requirements. I collaborated with stakeholders to determine AI prompting, executed the design direction, and conducted research to assess the feature’s impact.
OpenTable prioritized leveraging AI to help restaurants save time and create better content, starting with the technically straightforward and low-risk Review Management area, which I lead as the designer, and my team and I were tasked with building the new feature.
Based on the data above, qualitative data from previous restaurant interviews, and a brainstorming workshop that I led, we determined that restaurant review responses would be a good fit as an LLM feature because
We established that we would like the LLM to generate review responses that would make it easier for restaurants to respond to reviews. Through this feature, we would validate restaurants’ appetite for LLM-based tools.
What’s unique about designing a generative AI solution is that most of the effort wasn’t on the digital interface itself. Instead, I conducted discovery research and collaborated with stakeholders to define guiding principles and requirements for prompting the model, ensuring quality outputs and minimizing risks.
Across two beta rounds, we assessed the feature using engagement data, in-product feedback, restaurant interviews, and a diner survey of 368 participants. I led the restaurant interviews and survey, which revealed the following insights: