“AI fundamentally changes how we develop apps and what apps can do, and I would say we’re at the beginning of that revolution,” Intuit CTO Marianna Tessel said.
In a conversation with Nara Logics CEO Jana Eggers at Transform 2020 today, Tessel outlined some of the key ways that AI is changing the mindset at Intuit, with a focus on the app development process. Notably, Intuit is trying to adapt much in a similar way to how it adapted to the emergence of smartphones.
Mobile-first to AI-first
Transitioning from a legacy desktop-based software provider into a mobile-first technology titan was never an easy feat, but that is exactly what Intuit achieved. Founded back in 1983, the Mountain View, California-based company was renowned for its various PC-based financial applications including Quicken, TurboTax, and QuickBooks. With the advent of the modern smartphone era more than a decade ago, Intuit could have rested on its laurels and dismissed the emerging mobile app craze as a fad, but instead the company embraced this new form factor and went on to grow the value of its shares by more than 1,000% since 2008.
Today, Intuit is in the midst of another major revolution, as it moves to an “AI-first” ethos across its suite of apps. The company’s QuickBooks accounting app uses AI to automate repetitive work, automatically interpreting receipts and categorizing bank transactions, while it also uses machine learning to assess risk as part of its small businesses lending unit, QuickBooks Capital. Moreover, QuickBooks also ships with smarts such as cash flow planner, which is designed to help small businesses predict their daily cash flow over the next three months based on data such as recurring expenses and income.
Above: QuickBooks cash flow planner
Consequently, AI and automation are changing the end user experience, and no two customers will see exactly the same thing inside the app.
“This means that the experience of a user stops being static and [becomes] something that is dynamically generated — the user experience, we can now generate on the spot,” Tessel said. “You need to think about those many experiences that you can stitch together, so how you write the app is very different as a developer because you started from a very predictable flow to one that is personalized to the user and that is unpredictable.”
In short, rather than creating apps and then adding AI on at the end, apps will increasingly need to be designed from the get-go with AI firmly in mind. A good example here is the concept of the conversational user interface (CUI), which has risen to prominence thanks to the proliferation of chatbots and digital assistants such as Alexa and Siri — users are becoming more accustomed to navigating or seeking information using natural language, rather than wayfinding through tapping buttons or typing in keywords. It’s all about allowing users to get to wherever they want to go without having to jump through hoops.
“You might ask questions about your tax return. Well, in that context you need to be able to answer questions that might be from a different place in your app or in your domain,” Tessel said. “It’s a lot more dynamic, so how you write apps has to change.”
This change will likely be similar to what happened in the transition between desktop and mobile. Initially, software was built for Windows-based PCs or Macs, with the smartphone incarnation following after — this mindset was often reflected in the way that resources were allocated to each version of the app. By Tessel’s reckoning, app developers will have to adopt an “AI-first” mindset by default.