Machine learning algorithms aren’t new to QuickBooks, which has made increasing use of them over the past few years as part of its push for greater automation. However, Michael Hitchcock, Intuit’s head of product management for accounting and tax at QuickBooks, believes they can do much more. To that end, QuickBooks intends to have a particularly strong focus on artificial intelligence this year in service of a paradigm it has dubbed “all-in”.
Hitchcock said, by his own admission, that “All-In” is an ambitious agenda that acts more as an aspiration than a concrete goal. That’s what QuickBooks has called its overall strategy over the past few years of making incremental functional improvements to its ability to automate day-to-day accounting work, leveraging AI to identify what clients and accountants can focus on, and then providing the information needed to work more efficiently can.
“So when we say ‘all in,’ it’s sort of like that ambitious, borderline unattainable future state where someone can get into the QuickBooks ecosystem, connect their data, and start accounting,” Hitchcock said.
As much of accounting works today, he noted, there are manual forms that someone then connects to their bank or accounts payable department. Users have to go in and manually review each transaction to ensure the books have the correct categories, customers, and suppliers.
“The ‘all-in’ paradigm is, ‘Hey, how can we turn this on its head? How can we make it where we use the power of AI and automation to automate the busy work? … How can we We prepare the correct accounting journals based on the facts [our machine learning models] Have you seen this before?’” he said.
Hitchcock is confident that QuickBooks can achieve effectively “hands-free” accounting that can be performed with little to no human intervention. Part of this is due to improvements in the AI training model. Similar to humans, AI models need to be trained to understand what to do. Hitchcock said it used to take about six months to fully train a model before it could be used on customers. Now, not only can it be done almost instantly, but the training can also be done by the customer using the “Train QuickBooks” feature. Instead of spending months feeding data into the AI system, users now answer between five and ten questions about their business and needs. The clustering machine learning model does the rest.
“Today we learn the mode almost immediately. …You answer five to 10 questions and immediately train the model on how your business works,” he said.
Train QuickBooks was launched in the US market about a year ago and is currently being rolled out in other locations.
Hitchcock also heralded the forthcoming Guided Questions and Answers as another major factor in this push for AI. Just as Train QuickBooks is used for routine processes that don’t require much attention, Guided Q&A is suitable for complex situations that require human concentration. It is based on a computational knowledge engine and guides people through particularly tricky scenarios such as taking out a loan or accounting for something ambiguous. Hitchcock likened it to how Intuit’s TurboTax product works, except it’s at the transaction level.
“Rather than asking the client to, for example, create another chart of accounts to separate the principal of a loan from the interest, we ask a series of very simple questions a la TurboTax, just to be able to do proper accounting,” he said. “Think of it more like a simple button: I’m looking at this data in QuickBooks, this one looks hard, let me press the simple button.”
Hitchcock said that by having these two things working together, users can automate all of their busy work while also leveraging AI to take care of the critical matters that require their attention.
Additionally, he said that in a few months, QuickBooks will introduce an AI-guided end-of-month review feature that not only provides a summary of financial information and what tasks QuickBooks has already automated, but also curated to-do lists of the tasks that still need to be done.
As an added sweetener, he said this tool can also help clients provide better information to their accountants. The model can be set to flag not only on-balance-sheet transactions, but also low-context transactions. After marking, the program can prompt the user to provide additional information, which the accountant can then view.
“Every time we know an accountant sends an email or a text message and asks, ‘What’s up with that?’ We can fine-tune the AI engine to work for both this small business and this accountant, so that by the end of the month, when they get ready to prepare the finances, they have all the documents they need and all the contextual data they need need”, he called.
The month-end review is currently available to accountants and will be generally available to small businesses later this year.
Hitchcock believes such moves will align QuickBooks with where the future of accounting technology lies. Just as the profession itself has become increasingly intertwined with machine learning and artificial intelligence, so must the products they use.
During that time, issues of trust and control will rise in importance, he predicted, as these will pose the biggest AI challenge for the accounting industry. Unlike other applications such as imaging or writing, accountants generally require knowledge of what is going on in the system. They will not accept a black box without an audit trail. For that reason, he said, QuickBooks will aim to leverage automation, not to take control away from accountants, but to give them even more control.
“The challenge is that we look at AI as a black box,” Hitchcock said. “You give it an input, it spins for a second, and then it spits out an output. The big challenge we have is that we know accountants want ultimate control. They want to be able to harness the power of AI, but doing it that way builds trust. So that’s what we’re really focusing on. How do we make sure we control the accountants? How do we make sure they get visibility and feedback? How can they take a look under the hood so they can refine this because I want to give them confidence and make life easier for customers.”
