AI has the potential to transform internal business processes and products, but it’s not always easy to translate the potential of AI into tangible benefits. The data that drives AI is the most critical asset modern organizations possess, but its value is directly proportional to the number of people who can access, understand, trust, and derive insights from it. In the article below we summarize few learnings of Microsoft using AI in marketing, finance, and customer service.
Late payments prediction in finance
About 99% of Microsoft Customer transactions utilize some form of credit. Microsoft’s previous process was to contact 90% of customers with an email reminder about payments, but to reduce workload and improve customer experience, the company wanted to contact only those customers who are likely to pay late. The treasury and finance teams partnered with IT to create a model that predicts with over 80% accuracy whether customers are likely to pay late. The team used Azure Machine Learning Studio and a third-party algorithm called XGboost to create the predictive model. Microsoft segments the data, drawn from internal SQL Server data warehouse, by variables in a process called feature engineering. Engineers identify what data to use and then build a pipeline with data from the SQL Server data warehouse to enable the predictive model. Data scientists train the model using Azure Machine Learning Studio, then connect the data to the eXtreme gradient boosting (XGBoost) algorithm which creates decision trees. The scores go into SQL Server database and are displayed in Power BI reports to collections teams.
The AI-powered late payment prediction tool has reduced the number of customers contacted with payment reminders from 90 to 40 percent. In addition to predictions about specific customers, the company has also learned that complex invoices are more likely to be late, long-term, high-volume customers and partners are rarely late and can benefit significantly from payment automation.
Intelligent lead scoring and qualification in marketing
Microsoft’s marketing organization receives up to 10 million leads per year. The company was looking for a better way to score leads to reduce the amount of time sellers spend pursuing unproductive ones. Microsoft combined the marketing employees’ understanding of lead quality with the machine learning expertise of the company’s data scientists to create a lead scoring platform. The platform weighs thousands of variables to predict the probability that a lead will convert on any given sales channel.
To further qualify leads after they are scored, Microsoft created an AI-based lead qualification assistant called BEAM (Bot Enabled Augmented Marketing). The platform is built on open-source machine learning tools and Microsoft technologies including Microsoft Cognitive Services, Azure Machine Learning, and Azure ML Studio. It detects intent and context in customer emails to determine the likelihood that a customer is ready to purchase products or services. BEAM emails customers and evaluates their level of interest using natural language processing before they send them to sales. The AI model runs several machine learning algorithms to assign a value to each data point and generate an overall numeric score for the lead. The scored leads are then returned to the marketing engine through the API.
Microsoft’s AI-based lead scoring platform has helped the company more intelligently identify potential customers, ultimately improving conversion rates and marketing ROI. Using AI rather than traditional business rules helped the marketing employees streamline lead qualification and hand off fewer, higher-quality leads to sales teams.
Improvement of customer feedback analysis
Microsoft needed to find a solution to the huge amount of customer feedbacks it receives. To handle the problem the company created a sentiment analysis tool that makes it easier to interpret and act
upon post-transaction customer feedback. The tool analyzes sentiment to identify key factors that drive customer experience and breakpoints where the service didn’t live up to expectations. The company’s voice of the Customer team has several processes for acting on customer feedback, including support professional feedback loop and customer recovery feedback loop.
The AI-based customer feedback analysis starts with capturing and storing customer feedback from phone chats, email, and IVR surveys, then integrates data from 9 databases using SQL Server
Integration Services (SSIS). The solution processes customer feedback through machine learning workflows such as Microsoft Machine Learning Server on R language and Microsoft Translator. The insights are stored on SQL Server for teams to access, also stored through Power BI dashboards, and connected to Azure Dev Ops for coaching. Finally, these findings are surfaced through actionability
programs like Customer Recovery Feedback and Support Professional Feedback Loop.
The AI-powered customer feedback tool generates valuable insights in a fraction of the time, equipping the agents with the information they need to best serve the customers. For recovered customers, positive sentiment increased on average by 37% and Customer Satisfaction Score (CSAT) increased by 180% compared to their first interaction with Microsoft.
Effective cross-functional, multidisciplinary collaboration is key to long-term success. Softline as a Global Microsoft Partner has significant experience in AI-powered solutions. Ask for our experts’ counseling and gain all the benefit from transforming your business processes with AI. If you would like to read more examples in the field of sales and finance, download the free ebook here after registration.
Financial services providers are forced to rapidly accelerate the digitization of their businesses as managing data well is now the cornerstone for digital transformation. AI will become crucial in helping banks and others to automate operations, manage risk, prevent financial crime and cyberattacks, and personalize their services. Learn how AI is helping financial institutions to transform their operations to meet their current challenges and to prepare for future.
Enterprises are implementing a new generation of machine learning operations (MLOps) platforms that help them democratize and operationalize AI, allowing them to accelerate and govern the end-to-end machine learning (ML) model lifecycle. In this article we collect the drivers why organizations decide to develop Machine Learning model, and we summarize why Microsoft Azure Machine Learning is the solution.
Manufacturers globally are being disrupted and moving from making products to delivering products-as-a-service. To compete and grow, manufacturing companies need to shift their focus from engineering and production to customer outcomes. In this article we summarize those four key transformation topics that can help manufacturers stay competitive while meeting changing customer needs based on Microsoft’s ebook.
An essential building block of cloud modernization is a modern data and analytics platform because optimized data fuels modern operations. Data platform modernization is primarily focused on selecting the best cloud data platform technology to help your business achieve its strategic objectives and goals. To help you choose the best solution for your database modernization we provide a compact overview of the available Microsoft Azure SQL databases and their key features.