Intuit currently pairs A.I. with customer information already on file to tailor more specific financial advice and products for users. For example, Goodarzi said, Quickbooks can predict a small business’s projected cash flow and then recommend financial advisors that suit the firm’s performance. If A.I. predicts a strong quarter, Quickbooks might suggest a small business financial advisor that can help with expansion planning. If projections are less rosy, Quickbooks might direct the client to resources for securing additional business loans. 

Intuit does this for every one of its roughly 100 million customers, according to chief data officer Ashok Srivastava. “We’ve got millions of customers and we literally have millions of A.I. models running this operation on a daily basis to do that for people,” he said at Brainstorm A.I.  

A.I. allows “us to shift from what we used to have, which was a transactional platform…to a platform that’s smart, that is actually doing the work for you,” Goodarzi says. He hopes to transform Intuit from a financial toolbox to a sort of personal handyman. Therein lies the big revolution, Goodarzi says: with the proliferation of A.I., it’s no longer enough to help customers with a task. Instead, companies must complete the task based customers’ specific, individual needs. 

Intuit deploys A.I. across its own business operations in engineering, product design, and even in the workplace tools it uses like Slack. “We don’t have a group on the side working on A.I. It is core to everything that we design,” Goodarzi says. By incorporating A.I. into every nook and cranny of the company, Intuit is giving its employees an up-close look at what the technology can accomplish and how it can serve customers. 

Still, A.I. presents a number of ethical and legal quandaries for the company: How is customer data protected? Who has access to it? To safeguard customer data, Intuit uses the same privacy policies for A.I. that it employs for the rest of its business, Goodarzi said.

Newer A.I. products that aim to replicate human reasoning prompt even more nuanced questions: Can machine learning adopt human bias? Who should be accountable if it does? To address these thorny issues, Intuit established an A.I. Governance Committee, of which Srivastava is a member, that evaluates A.I. programs from the earliest stages of their development through to launch. The key element, Srivastava explains, is that Intuit doesn’t build its A.I. products and determine their ethics later. Rather, the company is aiming to incorporate ethics into the process from the outset, with the governance committee providing oversight.