I worked as the lead Designer at Bosh – Business Operations Self-Healing team.  

This team provides a SaaS based product that offers a comprehensive out-of-the-box AI engine to support automation in the business operations self-healing space from self-help, self-service to self-healing.  

So in short, they have many Machine Learning (ML) based use-cases that helps humans be better at their jobs, i.e, promote Intelligent Augmentation. 

Here, I worked closely with the Data Scientist (Tejaswini Kumar)  and Business Expert to design the Bosh Launchpad. Below are a few dashboards that shows the ML use-cases.  

There are 2 main personas that I had consider when designing this:  

• Business User  

• Data Scientist 

From the below launchpad, you can see the list of all the AI resolutions/use-cases, along with 3 main metrics – Model Quality, Resolution Usage (Number of times this use-cases was used) and the ML training metrics like accuracy, plus an additional retrain button.

Condensed View (business User)

The unexpanded/concise view is mainly targeted to a business user who just wants to see how the Resolution/Use-case is benefitting his business. But then a data scientist can expand this tile to get more specific information details.  

Expanded View (Data Scientist )

You can also see the different filters: Template type – Multiclass classification, multi-label classification, Regression that is currently offered by Bosh. This means that the solutions to these are 3 main ML problems that are currently provided at Bosh, and any current ML use-case falls under these 3 scenarios. Next there are filters based on the Model Quality metric, and the current status the model: active, inactive, terminated, failed, deleted. These filters should ease the use of overall launchpad usage, and to enrich the end-users overall experience.  

The question marks associated with each label/title contains precise information with examples on what these labels means.  

Below is an example about trust levels explanations

• Our first step into Augmentation: Ticket data
• To fix this problem, we introduced a user-case: Quality ticket creation
• This user-case would take our customer from 1 day to a week, but now using augmentation, which in this case is the ML models provided from the STI service, we send the tickets to the right people in a matter of second.
• We now have the ability to save both time and money for our customers.
• Forecasts Sales: Deploying AI-powered software can help you garner historical data bout the customer’s past purchases and help the sales department pull in better conclusions. It will also help you to know your least and most sought-after product. Further, it can forecast the products that can be promoted on a given date that can rev up your revenue bar. Things like these could have been surreal without the invasion of AI into our lives.