Meghna is a UX designer with more than 10 years of experience in managing all aspect of UX - research, design, testing and strategic projects.

AI in UX

AI in UX

 

AI’s ability to predict and behave as expected by the end users, was used to explore use-cases to align to user workflows.

situation

Having gone through multiple product evolutions over time, I have realised that it is time that our products evolve to the next level by acting intelligently. Current task flows feel very manual and time intensive. In order for the products to align to user workflows it would have to predict and behave as expected by the end users.

task

My job was to look at how AI features in our products evolution.

action

Players: Adding AI helps us track all the combinations of user actions and influence his journey. The AI can be given an end goal and it would figure out how to reach the end goal using the various reward system at various points in the player journey.

Operators: Allows operators to track and monitor user’s journey a lot more effectively. Allows them to see cause and effect. Allows them to experiment with a control group

Scenario 1: Influence player journey with promotions
Player journeys through products are not as predictable as we think. If we wish to provide rewards based on player actions, we cannot predict the different combinations of actions a user could take while using our products.

Scenario 2: Automatic player segmentation. 
Even though we have a concept of auto-grouping that allows players to move in and out of segments based on pre-defined rules and triggers, there is a need to take it to next level. AI allows creation of new segments on the fly based on patterns found in player behavior. It would look at player interest towards a promo and create micro-segments based on interest value and commonalities in the players.

Scenario 3: Chatbot for helpdesk
Most helpdesk queries are related to 4-5 primary issues. In order to resolve the queries each operator has a protocol that requires the agent to look through a series of info and conclude with a resolution. AI allows agent to identify an issues type and receive a summarised info needed to conclude a resolution. 

Scenario 4: AI for internal pages
Current search for internal pages are unyielding with random results in no serious order. With the help of ML, the search will improve by identifying the type of search query as a person, document, acronyms, definition and order it by relevance of date and importance. 

result

I have gone to create a promo bot using my research ideas from this project.

 
Competitor Research Framework

Competitor Research Framework

Style Guide

Style Guide