When Intuit sent AI agents to 3 million customers, 85% returned. The reason, according to the company’s executive vice president and CEO: Combining AI with human experience turned out to be more important than anyone expected, not less.
Marianna Tessel, executive vice president and CEO of the financial software company, calls this AI-HI combination a “massive request” from its customers and notes that it provides another level of trust.
“One of the fascinating things we learned is really the combination of human intelligence and artificial intelligence,” Tessel said in a new paper. VB Podcast Beyond the Pilot. “Sometimes it is the combination of AI and HI that provides better results.”
Chatbots alone are not the answer
Intuit, the parent company of QuickBooks, TurboTax, MailChimp and other widely used financial products, was one of the first major companies to bet on generative AI with its GenOS platform last June (long before they feared "SaaSpocalypse" (made SaaS companies have to rethink their strategies).
However, the company quickly recognized that chatbots alone were not the answer in business environments and opted for what it now calls intuitive intelligence. The dashboard-like platform features specialized AI agents for sales, tax, payroll, accounting and project management that users can interact with using natural language to gain insights into their data, automate tasks and generate reports.
Customers report that invoices are paid 90% in full and five days faster, and that manual work has been reduced by 30%. AI agents help close books, categorize transactions, run payroll, automate invoice reminders, and bring discrepancies to light.
For example, an Intuit customer discovered fraud after interacting with AI agents and asking questions about amounts that didn’t add up. “At first I thought, ‘Is that a mistake?’ And as he dug deeper, he discovered a very significant fraud,” Tessel said.
Why humans are still aware
Still, Intuit operates on the principle that humans are “always accessible,” Tessel said. The platforms are built so that users can ask questions of a human expert when they don’t get what they need from the AI agent or when they want a human to exchange ideas.
“I’m not talking about product experts,” Tessel said. “I’m talking about an accounting expert, a tax expert, or a payroll expert.”
The platform has also been created to suggest human involvement in “high-risk” decision-making scenarios. The AI reaches a certain level, then human experts review and rank the rest. According to Tessel, this provides a level of trust.
“In fact, we believe it becomes more necessary and more powerful at the right times,” he said. “The expert continues to offer unique things.”
The next step is to provide customers with the tools to perform next-generation tasks such as vibe coding, but with simple architectures to reduce the burden on customers. “What we’re testing is the idea that you can actually code without realizing that’s what you’re doing,” Tessel said.
For example, a merchant who runs a florist shop wants to make sure he has the right amount of inventory in stock for Mother’s Day. They can vibrate code an agent who analyzes sales from previous years and creates purchase orders when stock is low. That agent could then be ordered to automatically perform that task for future Mother’s Days and other major holidays.
Some users will be more sophisticated and want to be able to delve deeper into the technology. “But some just want to express what they want to happen,” Tessel said. “Because all they want to do is run their business.”
Listen to the full podcast to find out:
-
Why first-party data can create a "pit" for SaaS companies.
-
Why showing AI logic is more important than a polished interface.
-
Why 600,000 data points per customer change what AI can tell you about your business.
You can also listen and subscribe Beyond the pilot in Spotify, Apple or wherever you get your podcasts.
