Alphasense

Background

AlphaSense is a revolutionary AI-powered search engine for market intelligence used by financial firms and corporations across industries and geographies. With more than 1,000 enterprise clients, our mission is to enable knowledge professionals to acquire critical business insights and data with speed and conviction

My Role

I am the senior search designer at Alphasense, in charge of research, design & coordination of the main website. We had three more designers, for mobile, notebook, & marketing respectively. I worked on several projects namely: the earnings module, redesigning the entire search facets & filters, improving the Sisu design system that is used to build the elegent Alphasense web & mobile platforms, and worked on a testing framework for accessibility. I worked with several stakeholders including the Director of Product & SVP in New York, Scrum Manager in the Helsinki (Finland) team & the Front-End Engineering Manager in Pune (India) on a regular basis. In this case study, I will dive deeper into my latest project: How to use an AI-powered search engine to search for sentiment, cause, and reason in Earnings Transcripts, News & other docs for difficult questions that analysts ask. For instance, Why did Expedia's revenue drop in Q4 2019?

Why did Expedia's revenue drop in Q4 2019?

Problem

The business goal is to improve trial conversion and retention for Alphasense. The hypothesized problems are learnability and use case discoverability - users can’t always find the value-add potential of Alphasense. The strategy is to encode key client search use cases (where Alphasene can add value via search) so that users can’t miss them. A common use case category is corporate earnings (buy-side). And the new workflow for it is Earnings Preparation.

Purpose

Why are earnings workflows worth addressing? The earnings Season takes a lot of our buy-side analyst client’s lives. Earnings itself takes up around 20 weeks, which is 40% of the year. Preparing for the earnings takes up around 12 weeks, which is another 20% of the year. The idea generation phase/downtime is around 20 weeks, which is the rest of the remaining 40% of the year. Earnings phase is a time when efficiency is key, and resistance to learning new things is very strong. This is a big challenge and a big opportunity for Alphasense. The goals for this project are that it must be immediately obvious how to use Alphasense search for earnings. Alphasense must provide large efficiency gains, and enable digestion of the information in ways that aren’t possible elsewhere.

User Interview Script

In order to create a product to cater to the earnings, I've interviewed analysts about their workflow through the earnings phase. Below is the user interview script I created to conduct the interviews.

User Interview Script made for understanding the Earnings Workflow

Research Insights

After interviewing five analysts, I synthesized the research notes into the below three phases, namely: Earnings Prep, Earnings itself & Idea generation phase (downtime).

Research Notes from the Interviews summarized.

Earnings Workflow

Below is the user journey map I created for an analyst during the earnings preparation phase. With this workflow, I can determine when & how Alphasense can step in and make their cycle more efficient.

After conducting user research interviews, I developed the user workflows for the Earnings Preparation Phase.

Idea-gen Workflow

Below is the workflow for the Idea generation phase. As you can see there is a lot of multitasking involved here. There is a big potential for Alphasense to help analysts look for new changes in companies' earnings or news etc.

After conducting user research interviews, I developed the user workflows for the Idea Generation Phase.

UI Sketching

Brainstorming ideas for the Earnings module based on the research above.

Earnings Module V1

This is the first version of the Earnings module. When you search for smart synonyms related to earnings preparation, the earnings module will be at the top of your search results, showing the latest transcripts with the highlighted snippets related to your search keywords.

This is the first version of the small view of the Earnings Module that is shown in the search results.

This is the expanded view of the Earnings Module.

Prototype Interaction

This is a screenshot of the final interactive prototype done in Figma.