As Product Designer and Architect supporting on multiple design initiatives for client State Farm Insurance. Leading product design and delivery of AI and Machine Learning enabled Contact Center's Quality Management dashboard for State Farm Insurance.

Dashboard design and data visualizations for real time performance metrics and data analytics using AWS QuickSight and MS Excel. Supported multiple initiatives in the data visualizations for real time and historical metrics trend insights and analytics for oversight on Contact Center's operational efficiency and agent performance as well as in the Forecast of metrics and comparative analysis with historical data.

Note: The designs and deliverables could not be showcased due to non-disclosure agreement with State Farm Insurance.

Individual contributions included

• Leading the user experience design. 
• Information design, and UI design using Figma, Sketch and InVision.
• Data visualizations using  AWS QuickSight and MS Excel.

State Farm's Hercules Quality Management dashboard empowers it's Contact Center managers and supervisors to support agents and their teams with timely quality data and actionable insights to elevate agent performance and enhance customer experience.

Solution Highlights:

• Incorporates machine learning from Amazon's Contact Lens for detection and analysis of sentiment or speech behavior for voice contacts as well as for theme detection for real-time contact scenarios.
• Integrates Generative AI prompts with chat transcripts for empowering agents in responding to customer queries.
• Enables real-time monitoring of voice and chat contacts.
• Provides real time and historical metrics trend insights and analytics for oversight on State Farm's Contact Center's efficiency and agent performance for enabling identification of trends and data driven decisions to optimize team performance.
• Metrics tracked included Average Handle Time, Transfer rate, Adherence, CSAT score, Contacts handled, First Contact Resolution, amongst other metrics.
• Incorporates proficiency based routing.
• Integrated agent evaluations, balanced performance scorecard, and agent coaching
  

Workforce Analytics is a tool used in State Farm's Contact Center. It provides a way for visualizing actual queue performance, and comparing it to a forecast. It gives Operations team the ability to understand what's going on in State Farm's Contact Center business.

It uses forecasted volumes and forecasted goals, and creates a comparison between what was forecast to what is actually happening in the queues. It provides a real time view, either a daily view or a weekly view. Operations can select for different types of data groups such as Average Handle Time (AHT) and content volume, SLAs, and Average Speed of Answer.

It displays a graph with a comparison of the actual vs the forecasted data. This graph will crawl, depending on the time of day, and it allows Ops to get a visual comparison of the forecasts that was used, and the actual data that's coming in.

Forecast Management tool is an application on State Farm's Contact Center platform that is used to improve the ability to project the Contact Center’s operational performance and predict future Contact Volume, Average Handle Time, Full Time Equivalent, SLAs, Transfer rates, amongst other metrics. 

Key features include

• Using machine learning, it guides users through a list of selections, pre-processes data, and then creates a forecast of that demand that is used to create schedules
  

• The model that is used is Deep AR plus. It's a deep learning but auto regressive model that use for time series forecasting
• In June 2021, the accuracy of the model was around 80 to 90%. In Sept. 2021, the accuracies was around 85 to 95%
• Forecasters can make adjustments to the predictions. There are minimal changes that forecasters are doing now, and they're taking the ML forecasts and just going with that for their forecasting needs.

Multiple Queue Forecasting improves the time consuming process for forecasting for numerous queues every week

• Any number of queues can be included
• Generated forecasts can be exported individually or as a combined forecast
• Adjustments to the forecast can be made at the Q level

Multiple Week Forecasting improves the ability to project Contact Center’s operational performance in the upcoming weeks

• The application can forecast out up to 4 weeks
• What-if forecasts allows business to simulate outcomes with generated schedules
• Provides insights into how the contact centers expect to perform over the next 1-4 week.

Future State Metrics dashboard created for Operation's executives to monitor Contact Center performance with visuals intended to simplify analysis making it easy to consume information.

Business Solution:

• Displays a set of Contact Center and future state metrics, categorized by Voice and Chat data.
• Users selects one or multiple business segment to view Contact Center metrics that pertain to their business segment, or other business segments.
• Provides the ability to see data separated by business division, Contact Center's business segment and by associated Routing Profiles and Queues, and also by date range to view the metrics.

Business value:

• Reduced expenses, Reduced Full Time Equivalent (FTE), cost avoidance, etc.