INTELLIGENT SEARCH
Making Search more comprehensive and complete
OVERVIEW
Intelligent answers leverage the latest state of the art machine reading comprehension, to read and analyze billions of documents to understand the web and help you more quickly and confidently get the answers you need.
Bing now uses deep neural networks to validate answers by aggregating across multiple reputable sources, rather than just one, so you can feel more confident about the answer you’re getting. This was an exciting time for the team to do something really valuable for our users. I lead the design and strategy for the new set of “INTELLIGENT ANSWERS”
ROLE
Product Design Lead, Microsoft ・ 2018
Vision & Strategy ・ Product Design・ Concept Design・ Executive pitches and reviews ・ Fostering relationships between partner teams
Understanding User Expectations
As we started the project it was important to take a deeper look at the types of questions users have been asking and what their motivation and expectations could from a search engine or the web. The program managers helped in classifying different query patterns that users have been asking on Bing. The research team helped drive few online focus groups to understand what the expectations were as users asked those questions.
Perspectives from the Web
Just giving one perspective was the norm, with this approach we wanted to give users a holistic picture and bring both sides of the argument.
Get perspectives on thousands of questions
Compare entities on the web
A simple experience which would bring out the key differentiators for comparing two entities on the web.
Answers from multiple sources
This experience focused on bringing concise answers from multiple sources. This saved a lot of time for users as and provided an easy way to view insights.
Life is full of questions
Lot more templates that helped answer questions in a direct and concise way.
Impact
Impact of project was substantially high, as we made search more intelligent and helpful in common and complex questions. We saw increased engagement, drop in short clicks and repeat users.
