At Perplexity, pro search provides better sources, more accurate answers, and wider breadth of information. Yet, when compared to desktop, the mobile experience is lacking. With the goal to increase pro search signups in mobile, we can see large CLV and revenue gains over time given Perplexity's current retention of 85% (one of the highest in the industry. How might the experience improve while converting users to pro search?
Designed using the Perplexity brand guidelines I reverse engineered in Figma. A feature that allows users to explore personalized social content, reinforcing value and expanding relevant search topics.
I added an option to select different models, anticipating it will become essential as more models and granular searches are introduced. While automatic model selection enhances the current experience, this feature provides flexibility for advanced searches. The model icon and the name 'Tota 1.0'—Latin for 'whole' or 'entire'—reflect Perplexity's mission to redefine how information is organized, consumed, and created.
To test this idea, I would propose the following experiment:
Business Goal
User Goal
In order to validate this concept with end users, I propose a cohort based A/B test to daily active, non-paid mobile users to understand if personalized pro search converts at a higher rate than the control.
Using a product analytics tool like amplitude, I would measure the following to understand impact:
With the release of Anthropic’s ‘computer use’ feature 2 days ago, the move towards hands free interaction with AI looms overhead. Accordingly, what will this look like in mobile?
Using a simple use case, I’ve designed an interaction to explore how this might work in the Perplexity mobile app. Imagine a user interested in FAANG stock prices, a vertical Perplexity excels at. Using a takeover feature, how might a search that uses the reasoning model handle outside app communication and task latency? Below, I explore a simple version of an AI instructed to search for FAANG stock prices and post this information directly to the Apple notes app. Using push notifications, I simulate task latency and search completion.
A more advanced use case might be something like this:
A users tells the Perplexity AI to look up FAANG stock prices and log the price and details into their personal finance notion space. The user then asks perplexity to use Webull to buy those stock prices if the price per share is less than "X" amount. If Perplexity sees that this price is less than "X", then the user tells Perplexity to log this asset purchase into their Creditkarma account.
Imagine a world where user interfaces are no longer a barrier and AIs can use company APIs to communicate directly on behalf of the end user, accomplishing a collection of tasks that may have taken 10x more time for a person.