Useful Cognitive
Bringing an analysis of client feedback and competitive comparison to product and sales. This included machine learning assessment of hundreds of human comments with sentiment analysis and category discovery
Giving senior executives answers
BlueView started as a Watson enabled spreadsheet that categorized over 200,000 comments from clients coming from 26 different sources called Cognitive Feedback Analyzer (CFA).
In CFA the primary interaction is through a pivot table that takes over a minute to adjust the data with each selection
The primary users of CFA are analysts
I was challenged to take the information analysts have access to in CFA and make it suitable for Senior Executives. This is what became BlueView. BlueView’s purpose was to give senior executives answers to their first set of questions they typically look for without going to the analysts
This allows the executives to answer their first round of questions themselves, and have much better questions when they do ask the analysts for more details
AFTER
Simple clean design
Mobile Deployment
Easy to Use
Answers the most common questions
BEFORE
Excel pivot table
Very slow
Excel skills needed
Frustrating to use
In BlueView I transformed the data from the CFA spreadsheet to a mobile app targeting the iPhone
BlueView is centered on NPS scores from clients but it also enhances NPS with a new score: NSS or Net Sentiment Score
NSS is calculated similarly to NPS with a scale from -100 to 100, but is based on the sentiment of actual comments from customers
The comments from customers are processed by Watson to give a sentiment score (positive, neutral, negative) and a category based on standard categories from the Client Advocacy Office (CAO)
Users can find details about the client they are interested and drill all the way down to a specific comment
Over time, more data sources are being added to Blue View
BlueView Real Answers
Having real client comments easily accessible will make understanding what to do with an NPS score much easier. Studies have found that NPS is difficult to act on, but the comments that come from the standard Likelihood to Recommend question that NPS is based on are highly valuable
NSS extends the concept of NPS to a much larger more general set of comments. This brings many more datasets into the hands of executives as they are making decisions
As the designer, it was important for me to emphasize the words our clients provided. The Heatmap shows a summary of client comments categorized and rated by sentiment. Decision makers can quickly see what categories have the most comments and the NSS score for each, so they know the prevailing sentiment.
My design transformed a complex spread sheet into an easily consumed mobile application allowing our executives to get answers about what our most important clients are thinking about IBM.
