Back to Hub

The Rise of Conversational UI: How Chatbots are Used in Purchase to Pay

11/27/2018 By

“In a survey by Oracle of chief marketing officers, chief strategy officers, senior marketers and senior sales executives from France, the Netherlands, South Africa and the UK, 80% of the respondents said they already used chatbots or planned to use them by 2020.”1

Interaction between humans and computers has progressed rapidly in recent decades; I almost laugh looking back at how recent it was that we did not even have mobile devices and remembering the clunky first versions that did not have anything close to the functionality of smartphones today. But as we’re all hyper aware now — computers and smart devices are ingrained in how we live, work and play, so it’s no surprise that our relationship with technology is going to the next level with innovations like virtual assistants and chatbots. But first, a little history lesson on how user interfaces (UI) have evolved.

The History of UX/UI

The very early days saw punch cards and other media that users would insert into a computer to input data, offering no real-time interaction. Next came command line interfaces, (where you typed commands in the command prompt window in green text to tell the computer what do to — remember that?) which allowed a little more real-time interaction. Finally, came the graphical user interface (GUI) and pointing devices that completely changed the game and provided powerful user experiences (UX) — think icon menus like Microsoft Windows or how apps are displayed on the Apple iPhone. Now UI/UX designers are setting trends for web application UIs using flat graphics (not 3-D) and minimalistic designs.

The goal of all this progress has been to give people a better user experience, enabling them to interact with the machine in the easiest, most intuitive way possible. And what’s more intuitive than chatting? Enter the chatbot.

What’s Behind the Curtain

Machine learning and artificial intelligence (AI) are powerful technologies adding tremendous functionality to business software systems and a lot of value to users. Experienced users will commonly associate these technologies to large data sets with predictive and intelligent decision-making using that data. But a huge focus area for computer scientists working on AI algorithms is natural language processing (NLP), or the ability for a computer system to effectively process human language. With powerful neural networks and deep learning algorithms, NLP has made progress in leaps and bounds in the last few years. We’re not talking about those awful automated voice menus on help lines anymore — we’re talking computers that can listen, understand and get smarter as they communicate with humans. Can this technology literally allow us to realize the dream of “the best interface is no interface”? We say yes (for the right use cases)!

Chatbots in Purchase to Pay

With the proliferation of devices such as Amazon Echo and Google Home, artificial intelligence has become a part of life, whether we realize it or not. These devices were built from the ground up with conversational interfaces only.

Purchase-to-pay software systems are a little different. Entering an invoice using a conversational interface will probably not work. But many functions can be achieved using a chatbot for purchase to pay. Imagine the following real-world situation:

You are a new employee who needs to mark your new laptop as “received” in the purchase-to-pay solution. With a chatbot, this can be done in a matter of seconds:

  1. You log in to the purchase-to-pay system and a virtual assistant (a chatbot) greets you and asks how it can help — it’s just like the friendly assistant that you always wished you had, but your job grade never allowed! You ask it to find the PO from awhile back with the laptop that was ordered for you — just as you would chat with another human.
  2. The assistant quickly finds the PO and already knows the laptop was for you and asks if you want to receive the PO. You answer, “yeah” and it receives the PO for you. This assistant totally gets you! You thank it, sign out of the purchase-to-pay system and get on with learning how to perform your new role on your new laptop.

This is just one example of how chatbots can be used in purchase to pay. Think of all the possibilities where tasks could be completed by simply asking a chatbot for help — no more searching, scrolling, clicking or typing — just simple questions with accurate answers. Need to re-order supplies: Simply tell the chatbot. You need look up a previous invoice to see how much a contractor billed you: Simply ask the chatbot. The possibilities are endless.

Game-Changers for Purchase to Pay

“According to a study by Stanford computer scientist James Landay, speech recognition is on average three times as fast as typing on a cellphone, and the error rate has dropped 3.6%.”Chatbots save time, make the purchase-to-pay solution easy to use, eliminate the need for users to go back and forth with questions to one another and reduce frustration. Ultimately, companies reap unparalleled efficiency and garner high levels of user adoption for purchase-to-pay systems while delighting their users.

How Basware is Using Chatbots

Basware’s new virtual assistant, driven by artificial intelligence, was released May 18. The Basware Assistant is here to help with day-to-day tasks. There will always be functions in the system that require traditional UIs or RPA to accomplish, but so many use cases can be achieved simply by using a chatbot that is NLP ready and intelligent about the purchase-to-pay domain as well as the application. The Assistant is exactly that — and we can’t wait for you to start chatting with it!

Learn More

To learn more about Basware Assistant, download the factsheet, and as always contact us — we’re here to help!


Pramod Prabhu is Basware’s director of R&D and chief architect.