Technology is getting faster and more complex in banking and other industries. Forget operations that rely on a standard server to standard desktop with software updates delivered every year or two. Now the core data and apps can be on mainframes, server farms, in the cloud or delivered through APIs with IoT on the edge. And the user can be using a desktop, notebook, tablet, phone or watch running MacOS, Microsoft, Linux, iOS and various flavors of Android.
The problem is that banks have modernized the way they test and modify software by looking at it through the eyes of the customer, he said.
We have software that can use any app or any web site as a user would. It has an AI engine that can create users and imbue them with a personality and then watch as they use the banking systems and explore, following their user journeys and finding problems.
If you are looking at underlying registers data it might look fine, but if you use it as a customer does, you find the problem, Bates said.
Eggplant combines testing and monitoring so firms can continue to test a product when it is operational. With continuous delivery, discerning customers and intense competition, automated testing and monitoring are necessary people are too slow and can't handle the number of endpoints that are vital in financial services.
It could be a performance problem caused by a third party and with a large embedded graphic that takes 60 seconds to load on mobile.
Millennials get bored after three seconds, so off they go. Or the presentation might have bad colors that are harder to read on one phone or one browser than they are on another. You have to test every permutation. A tool kit might obscure a key piece of information on a single platform its all that kind of weird stuff that you need to be able to test really quickly.
To excel banks need to develop what Bates calls biz-dev-ops structures that bring together the business, developers, and operations to create a continuous loop of testing, monitoring combined with reporting at the business level.
What do we know about their failures? Not a lot. All they said was that a long-planned system update went wrong, and then they fixed it. The thing that interested me in TDs case was they said they had fixed it, the customers said they hadn't, and the customer spoke through social media, he said.
See the Down Detector for a view of prominent companies whose systems have crashed.
Eggplant can understand the images on the screen and the text and can use a remote control to access an app through a keyboard or mouse, just like a human operator.
Nobody has brought together testing and monitoring and doing it through the eyes of the customer, Bates said.