Unit4 provides enterprise resource planning systems to governments, not-for-profits, higher education and professional service firms. Based in Utrecht, the Netherlands, it’s known in corporate travel for powering Sabre’s Central Command travel agency back-office software. It also offers an expense management product and, as of this spring, a digital assistant. Named Wanda, the chatbot stems from Unit4’s focus on “self-driving applications.”
The idea is to let human users easily interact with the company’s enterprise software components. Wanda is accessible via popular messenger apps and talks to other digital assistants. Formally announced this month, it is starting with five features. In addition to purchasing, task and time management, one of them is a travel assistant.
It doesn’t make travel reservations. Rather, users tell Wanda (by typing or speaking) about a business trip they’re planning and Wanda asks the user’s manager for approval.
It’s conversational. During a demo for The Company Dime last week, Unit4 global head of presales and product strategist Thomas Staven told Wanda that he needed to travel to Stockholm on the coming Wednesday. Wanda asked for the trip purpose and cost estimate. Users can input an estimate or let Wanda come up with one based on similar, past trips.
Wanda reports back with approval or rejection. If approved, the user goes off to book travel in some other channel.
In a future version, Wanda might search for the best flights and hotel rates — perhaps based on previous trips — and handle the booking via the client’s travel management company. Staven said the company is closing in on embedding travel policies and letting Wanda answer related questions.
A next step is integration with Microsoft Office 365. In that way Wanda would be proactive. When finding calendar entries related to an upcoming business trip, it would ask if the user wants to send a travel request.
The digital assistant also helps with expense reporting. Users can upload a receipt image or manually enter the expense item’s type, amount, currency and date. Wanda guesses to which expense report that item belongs, allowing the user to adjust. Depending on configuration it also may ask the project with which the expense should be associated. It might also press the user on policy compliance. In the demo, after Staven uploaded a taxi receipt image, Wanda asked why he didn’t use public transportation.
Wanda’s roadmap includes location services, more policy features, importing receipts embedded in emails and integration with new payment mechanisms.
It’s built on the Microsoft Azure cloud platform and uses natural language processing, machine learning and artificial intelligence.
Wanda doesn’t have its own voice. “We decided to leverage apps already in use,” Staven said.
It’s accessible through such popular chat apps as Slack, Skype and Facebook Messenger, with more such connections to come. Unit4 introduced Wanda to Amazon’s Alexa.
“This ability to bring natural language controls to any app is a massive game changer for enterprise computing,” according to a January 2017 blog post by Unit4 chief architect Claus Jepsen. He added that it is “great that we can use this for booking travel or claiming expenses for example, but more crucially the ability to ask questions and follow-up in the way you’d have a normal conversation is incredibly powerful. For the first time, speech recognition available through applications like Alexa gives us the ability to build conversational applications that business customers can and want to use every day.”
The company also wants Wanda to befriend Microsoft’s Cortana, Google Home and other digital assistants.
Wanda users must work for companies that have installed Unit4’s expense management software. It comes in two flavors. Business World Expenses is part of the larger Unit4 Business World ERP offering. The company calls this the “standard solution.” Travel & Expenses is the “premium” standalone system designed “for the most advanced requirements.”
Like a lot of systems, Travel & Expenses features receipt capture.
“The innovation is what happens afterwards,” said Unit4 global product manager for time and expense Oistein Nordhagen.
By that he means the system predicting the type of expense and determining if the account has a preferred relationship with the supplier. With machine learning, it gets better at it. Because it’s “public cloud learning,” Staven added, the system learns across all system users, not just among those in the same company.
The newest module now in testing includes “travel management connectors.” Nordhagen said “open APIs” could show expenses that are in and out of policy or whether the user booked directly or with the TMC. Depending on how client companies configure the system, end users might have to explain out-of-policy items before being able to submit.
In North America, Unit4 resells Travel & Expenses via ServeVita, a software integration consultancy.