Intuit’s strategy is to be an “AI-driven expert platform” where we solve the most pressing customer problems and deliver wonderful experiences using AI.¹ To that end, and to ensure our innovation is customer driven, Intuit® follows its “Customer Driven Innovation” (CDI) principles.²
In this post, I will share some powerful examples of how Intuit’s AI strategy works with its CDI principles to create customer driven AI solutions.
Principle 1: Find the problem
When we set out to design an AI solution, we start with the customer problem. The customer problem must be significant enough to justify the investment and must obviously be one that hasn’t been solved yet.
Principle 2: Ensure the tools are right for the job
Even though AI is powerful, it cannot solve everything, especially when you consider data availability and performance limitations. We seek out customer problems that we can solve well using AI, with the data we have at hand, and with high performance.
Principle 3: Create the advantage
We focus on areas where the more customers benefit from the AI solution, the better the solution becomes. For example, an AI solution that we can iteratively improve, as we get more data from real customers’ interactions, can result in a durable competitive advantage.
In the next sections, I will show how Intuit follows CDI to solve important customer problems, using the AI capabilities of two of its products: TurboTax and Mint.
Expediting the Refund Experience for TurboTax Customers³ ⁴
TurboTax is the #1 best-selling tax preparation software to file taxes online. As a result, TurboTax processes a huge amount of tax refunds.
In order to receive the tax refund, the TurboTax user has to provide their bank account details. However, 1% of TurboTax users mistype their bank account details, which delays their refund by eight weeks!
This sounds like a major customer problem, but let’s go over the CDI guidelines to make sure it meets our criteria for a worthwhile investment:
#1: “An important, unsolved customer problem …”
1% of users mistyping their bank account details, resulting in an eight-week delay of their funds, is definitely an important, unsolved customer problem.
#2: “… that we, and those we enable, can solve well …”
We could solve this problem, and significantly reduce the probability of mistyping bank account details by asking the customer to verify their details when the AI model indicates it might be a wrong number.
#3: “… and build a durable competitive advantage.”
AI gets better with more data. As more customers use the solution, we gain more data, allowing us to improve it even further.
To solve this challenge, we applied an AI deep-learning algorithm to reduce payment failures. The AI model, called Verum (“valid” in Latin), reverse-engineers bank account number patterns to predict the likelihood of each account number being invalid.
This resulted in instant notifications of this possible error and reduced invalid account number issues by 55%, preventing the delay of more than $140 million in tax refunds in just one tax season.
How Mint Is Helping Users Prevent Overdraft Fees⁵
An overdraft fee can be a headache and often is one of the most expensive fees from a bank. There’s a good chance that at some point in the last year, you have seen the dreaded “NSF Fee” transaction on your bank statement.
If you are unfamiliar with the term, NSF, or Non-Sufficient Funds, is when a customer tries to make a purchase without enough funds in their bank account to cover the transaction. The amount overdrawn incurs interest at a rate determined by the account provider and likely some additional (hefty) fees. For example, if you overdraw your bank account by a minimal amount, even by $1, you may be on the hook for a $30 (or more!) NSF fee per the terms you agreed to when you signed up for your account.
This can really add up. Every year, U.S. consumers pay more than $15 billion in overdraft fees. On Mint, we were able to see that $250 million in NSF fees are paid out annually to the top five banks alone.
NSF fees present a big problem. Let’s go over the CDI guidelines to see if it meets our criteria:
#1: “An important, unsolved customer problem…”
$250 million in NSF fees is a major customer problem, especially when you consider that customers are being charged at a time when they have a monetary shortage. In addition, this challenge hasn’t been solved before.
#2: “… that we, and those we enable, can solve well…”
We can solve this problem, since NSF is somewhat predictable when you have the customer’s historical financial data. For example, if there is a specific date when money leaves the account, and there is usually not enough funds in the account to cover it, there is a high chance that an NSF fee will be charged on this date. Knowing in advance that an NSF event might occur can help the customer avoid it.
#3: “… and build a durable competitive advantage.”
This solution has enormous potential, and that potential grows over time. As more customers interact with the solution, we can learn how to make it better.
The result was fantastic: We created an overdraft prediction service for Mint users, a smart technology leveraging artificial intelligence and machine learning to predict when an overdraft is likely to happen, based on spending patterns. The service saves Minters more than $900,000 a year.
Winning AI solutions
AI is extremely powerful, and it enables solutions we couldn’t have imagined possible in the past. When we at Intuit combine AI with clear CDI guidelines, we’re able to provide our customers with the best possible experience.
1. Intuit, Intuit®: Our Strategy. (n.d.). Intuit®. (n.d.). Retrieved May 25, 2022, from https://www.intuit.com/strategy/
2. Intuit, CUSTOMER DRIVEN INNOVATION — Intuit Labs. (n.d.). Intuit®labs. Retrieved May 25, 2022, from http://www.intuitlabs.com/customer-driven-innovation
3. Mintz, I., & Intuit Developer Team. (2021, February 25). Expediting the refund experience for TurboTax customers. Intuit Developer Community Blog. Retrieved May 25, 2022, from https://blogs.intuit.com/blog/2021/02/25/expediting-the-refund-experience-for-turbotax-customers/
4. Mintz, I. (2021, May 2). Ido Mintz, Intuit Israel: AI for the Masses: Avoiding $500M of Bank Account Mistakes. YouTube. Retrieved May 25, 2022, from https://www.youtube.com/watch?v=Y-Ycokqc9XQ&t=11s
5. Mint. (2019, April 16). How Mint is Helping Users Prevent Overdraft Fees. MintLife Blog. Retrieved May 25, 2022, from https://mint.intuit.com/blog/updates/how-mint-is-helping-users-prevent-overdraft-fees/