<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=552770&amp;fmt=gif">


Desired Capabilities

  • Customer Analytics and Insights
  • Improved Regulatory Compliance
  • Data Visualization
  • Text Analytics
  • Operational Efficiency

The Situation

With increased regulatory focus from the CFPB and the OCC, a Top 10 bank needed to build a compliant, customer-centric complaints management program. They needed repeatable processes to consistently define, capture, organize, analyze, disseminate and act upon customer complaints data across the enterprise — and ultimately to improve customer experience, lessen the burden of consistent categorization for phone associates, proactively identify potential regulatory issues, and spark insights necessary to reduce complaints.

The Challenge

Address customer and regulatory expectations by improving complaint definitions; front-line training; and complaint identification, data aggregation, research and resolution processes.

Meaningful Outcomes

  • Developed a single functional complaint definition across the enterprise.
  • Created an enterprise data repository and built a robust text analytics model. 
  • Launched interactive reporting focused on insight generation by highlighting hot spots, emerging trends and the ability to pull complaints based on their sentiment.
  • Implemented robust processes to share prioritized, actionable insights with specific leaders accountable for driving and reporting back on actions. 
  • Tracked complaint volume trending and severity to ensure actions taken would drive the intended impacts. 
  • Reduced response time to pull data for different events and regulatory requests from a week to less than four hours. 
  • Enabled an integrated view across all types of customer feedback that expanded beyond complaints.
  • Generated accolades for the bank from regulators on the level of detail and data quality built into the new complaints program.

Our Approach

Our team first ensured that the bank’s definition of a customer complaint aligned with regulator, industry and internal expectations, and that every front-line associate understood how to consistently identify and escalate complaints. Building on that foundation, we sourced and aggregated complaint datasets from across the company into an enterprise repository to enable reporting and analytics.

Once the data was available, our analysts built a text-analytics model to categorize records by complaint type and regulatory risk, ensuring improved accuracy throughout the review and training stages. We developed a reporting suite to provide actionable insights for senior leaders and front-line associates, enabling them to easily focus on complaints relevant to their area. We also worked with leaders to establish monthly routines for reviewing materials with their teams and built-in processes to track and hold teams accountable for conducting root cause analysis and driving valuable improvements.

With that solid foundation, natural language processing algorithms were trained to automate the process of classification and learning, which could be performed far more easily than with human intervention. These tools could be used to identify key words and phrases for further human review, analyze and classify sentiment to meet the immediate need, and uncover patterns in and trends from agent conversations for future training, policies and to derive new behavioral insights about the customer base.