The Power of Machine Learning to Travel, Expense, and Invoice Solutions

Machine learning (ML) holds such promise for improving job satisfaction and reducing errors and costs for corporations. At SAP Concur, we use ML to make everything from travel booking to expense auditing smarter, more automated, and easier for employees.

We have a long history of using machine learning to add intelligence to applications and eliminate manual processes. Unlike many companies, we strive to reduce the time users spend in our application. At times that means bringing the task to the user and other times that can be achieved through automation. ML presents exciting opportunities for SAP Concur to do more on behalf of our users, aligning the employee’s need for simplicity with the company’s need to manage risk, compliance, and achieve cost savings.

SAP Concur was founded on the question of how to have an expense report write itself, and we are getting close. Today you can take a picture of a receipt and it is automatically added to your expense report with the right amounts, details, and classifications. When you book travel with SAP Concur, your itinerary is automatically added to your expense report ready to be submitted once you touch down at the end of your trip. We also have intelligent tools built with machine learning so that receipts are added to your expense report right from your inbox. And we are working on many more innovations to get us to a point where the expense report truly will write itself.

As part of SAP, we’re proud to contribute to SAP Leonardo machine learning capabilities, and leverage expertise from Leonardo to continue to innovate and improve our products and services. And we’re excited to continue to broaden our ML knowledge across the company through close collaboration with the SAP Leonardo team.

Here are a few more ways we’re using machine learning today, and thinking about how it can impact the future:

Expense and Invoice

Expense reports and invoice processing today often require a lot of manual data entry. We’re using machine learning to help reduce time on task so employees can spend less time on expense reports, and more time adding value to the business.

Concur Invoice uses SAP Leonardo to extract data from invoice images and classify data, helping customers process more invoices and increase accuracy, while reducing technology and operational costs. The next step is to continue to research how ML can improve the overall invoice management process.

Concur Expense uses machine learning technologies in conjunction with Office 365 to create expense report entries directly from an emailed receipt that hits their inbox. Concur Expense on Outlook is currently available in beta in the U.S. and Canada.

ExpenseIt goes beyond simple OCR by using machine learning to automatically turn photos of receipts into completed expense reports for end users. We’re able to determine the location and vendor of an expense more often than that information appears on a receipt.

Concur Detect by AppZen applies machine learning to audit 100% of expense reports, before reimbursement occurs. Using artificial intelligence and ML, Concur Detect by AppZen validates whether expenses are legitimate and accurate by cross-checking them within seconds against hundreds of external data sources. Concur Detect by AppZen is integrated with Concur Expense, enabling seamless, automatic processing of low-risk expenses with high-risk reports flagged on the dashboard for review and alerts sent real-time.

Travel

TripIt gives users a beautiful mobile experience by instantly organizing emails from travel bookings. TripIt uses machine learning to find opportunities to enhance the way it organizes travel plans from emails, allowing it to scale to support travelers and their plans around the globe.

Hello Hipmunk, a free virtual travel planning assistant, uses Natural Language Processing and AI-powered travel search to provide travel advice and recommendations via a travel bot.

Concur Locate is designed to help organizations perform Duty of Care functions. It uses ML to predict where employees may be via credit card data, itineraries and other sources, enabling companies to reach all employees, not just travelers and VIPs, in the event of an emergency.  

Concur Mobile uses machine learning to provide better hotel recommendations, based on hotels previously booked by the user, as well as coworkers. Over time, this becomes more accurate and we see an increase in a user booking hotels which were recommended.

What's Next?

By unshackling smart people from laborious menial tasks, machine learning can help unlock new opportunities for interesting, creative, strategic work — the type of meaningful work that helps both employees and businesses get ahead.

Stay tuned for a deep dive on how we’re pushing the boundaries of machine learning in travel, expense, and invoice through vast amounts of data, collaboration with the SAP Leonardo team, and evangelism within SAP Concur. And be sure to check out our whitepaper on machine learning and the value to accounting and finance.