Skip to content
Microsoft in Business Blogs

Change is no easy task. As with most new technologies, AI adoption brings its share of cost and risk as enterprise leaders and department leads work together to build out their experience. Merely identifying an entry point has brought plenty of AI projects to a halt. Since early 2019, I’ve worked with dozens of IT innovation teams and business leaders at banking, capital markets, and insurance organizations to help guide their AI adoption efforts. These financial services organizations eagerly want to use AI to augment their productivity, reduce costs and minimize human error, but frequently struggle with how to start their implementation. In that time, I’ve identified three AI scenarios that are most frequently requested by these business leaders:

  • Forms recognition/OCR
  • Call center automation
  • Text analytics

Reduce repetitive work with forms recognition
The first and most popular use case is for large enterprises inundated with documents that rely on human labor to extract and process the information in those forms. These documents come in all varieties — customer account forms, invoices, PDF’s, or even insurance claim documents. As a result, they require an AI solution that:

  • Performs optical character recognition (OCR) on the document for both printed and handwritten text
  • Extracts specific fields such as name, address, account number, etc.
  • Takes that extracted information and feeds it into some back office system for further processing

The prescriptive AI solution we’ve built for document management consists of OCR and other “vision” related AI products that scan the documents and convert those pixels into text. This process is followed by another AI tool that extracts the required fields — typically a “box” that reads name, address, account number, etc. – from the form itself. The extracted information is then fed to a Robotic Process Automation (RPA) system that sends the information to some back office system. Finally, the back office system performs an action on this new data like issuing a check for an insurance claim.

Obtain valuable insight with call center automation and text analytics

Much like the need for documentation, call centers play a huge role in corporate accountability and maintaining a positive customer experience. Every day, thousands of service representatives engage millions of customers to ensure their expectations are being met. For most organizations, this means recording and logging conversations and tracking resolutions. Call center automation uses AI to go beyond prerecorded service lines and helps enterprises:

  • Transcribe recordings into text
  • Perform sentiment analysis
  • Translate language as necessary
  • Analyze recordings

Text analytics is an excellent partner for automated call center transcriptions. The option to use AI as a resource for developing insights from documents and legal papers gives growing organizations a considerable productivity advantage. Instead of digging through call logs, analysts can focus their efforts on making more valuable recommendations for the business.

For call centers, text analytics are mostly being used to observe agent performance and analyze customer sentiment around an organization’s products and services. We’ve also seen text analytics being used to extract information from call center recordings to spotlight trending issues happening on a certain day, week, month, etc.

This immediate access to translation, sentiment, and analysis delivers incredible value for finance houses. By extending the knowledge capture abilities of the call center beyond the traditional CRM, call center automation generates valuable data that organizations can use to redefine their product offerings, adjust messaging, and more.

Start your AI journey

Don’t let AI feel like an inaccessible undertaking. At Microsoft, we’ve developed pre-built AI models for forms recognition, call center transcriptions, and text analytics to provide accessible automation solutions without the need for a large team of data scientists.

The road to automation is far from insurmountable. While a handful of companies have used custom AI models for over a decade, the introduction of these pre-built models for vision, speech, and text analytics show an attractive ROI and are a great way for enterprises to get started on their AI journey. You can find these solutions and more at Microsoft’s Cognitive Services Suite, or reach out to me on LinkedIn to start a conversation about how we can bring AI to your organization.