For over a decade, Intuit has quietly been getting ready for a generative AI revolution that a big a part of the world solely started to grasp within the final couple of years.
“We had been already engaged on generative AI capabilities when the remainder of the world began investing closely in it,” Ashok Srivastava, chief AI and knowledge officer at Intuit, instructed AIM. “We had seen the advantages and had been able to go.”
Srivastava defined that this early funding culminated within the creation of Intuit’s GenOS, launched in June 2023. GenOS now powers all of Intuit’s generative AI and agentic experiences, abstracting away the complexity of varied underlying programs—knowledge platforms, consumer interfaces, and back-end infrastructure—to permit for large-scale deployment of AI brokers.
“We created GenOS to keep away from the inefficiency of getting hundreds of builders constructing hundreds of separate AI functions,” Srivastava mentioned. “That might imply shedding the facility of the platform we’ve constructed.”
Quite than constructing fragmented instruments, GenOS permits Intuit to ship constant, scalable AI experiences by integrating analysis, personalisation, safety, and fraud prevention immediately into the core platform.
The system now powers production-ready AI brokers, together with accounts receivable and accounts payable brokers designed to automate money circulate administration duties. These brokers have been alive since September 2023.
Why Agentic AI
Srivastava brings a deep background in AI and machine studying, having labored in neural networks, sign processing, and management programs since his graduate faculty days. He’s authored over 100 papers within the subject and continues to contribute as each a pacesetter and a technologist.
“Did I do know it was going to be precisely what we noticed in ChatGPT? No. However did I do know we wanted to spend money on AI as a core functionality? Completely. And we did that.”
Whereas many firms had been racing to construct AI integrations, Intuit took a distinct path by creating an working system from the bottom up. This gave engineers and builders a unified framework that would ship constant, dependable, and safe generative AI experiences.
Microsoft and Google are pouring billions into their very own LLM ecosystems. Srivastava emphasised each platform depth and knowledge richness, for instance, Microsoft simply onboarded round 11,000+ AI fashions on its Azure AI Foundry.
However he’s extra targeted on outcomes than sheer quantity. “Do we’d like 11,000 fashions to get our prospects paid quicker? No. In all probability, we don’t—as a result of we’ve already achieved it.”
Intuit’s agentic system can get prospects paid 45% quicker, a mean of 5 days sooner, because of automated transaction matching and bill evaluate. That type of tangible final result, Srivastava says, issues greater than the underlying mannequin depend.
“If I’m operating a small enterprise and I receives a commission quicker, that’s a win. I wouldn’t care if there have been one or 11,000 fashions behind it.”
“We didn’t need a whole lot of disconnected instruments. We needed a single system that builders can construct on, with safety and consistency,” Srivastava mentioned.
GenOS permits seamless integration of over a dozen LLMs—from industrial choices to open-source fashions fine-tuned in-house. This flexibility permits builders to decide on the most effective mannequin based mostly on safety, threat, and efficiency, all with out compromising consumer expertise.
“We’ve constructed analysis providers that work throughout all our agentic experiences. Should you don’t try this, some experiences work, and others don’t—it turns into inconsistent. Our working system ensures every little thing scales.”
On the information facet, Intuit’s entry to wealthy, structured monetary and behavioural knowledge permits the corporate to personalise AI experiences in real-time, delivering money circulate forecasts, clever suggestions, and context-aware automation tailor-made to the client’s quick wants.
“Our experiences are tuned to the second. That offers us a critical benefit.”
Open Supply as a Strategic Pillar
One other pillar of Intuit’s technique is its dedication to open supply. Tasks like Admiral, NumaProj, and Agroproj, are a part of the corporate’s effort to contribute to the broader neighborhood and leverage the most effective accessible applied sciences.
“We care and provides again—it’s one in every of our core values,” Srivastava mentioned. “We wish to construct instruments based mostly on our experience so others can profit.”
Intuit has acquired the Finish Person Award from the Cloud Native Computing Basis twice—in 2019 and 2022—an honour reserved for firms that contribute and scale open supply applied sciences successfully.
Whereas Srivastava declined to select a “favorite youngster” amongst Intuit’s open-source initiatives, he famous that NumaProj has delivered important worth each internally and externally.
Intuit’s platform additionally powers a set of widely-used merchandise together with QuickBooks, TurboTax, Mailchimp, and Credit score Karma. Every is designed not simply to perform as a instrument however to offer “done-for-you” experiences, the place AI does the heavy lifting and the consumer solely intervenes when crucial.
“What retains me up at night time is considering how we are able to truly assist small companies and shoppers do higher,” mentioned Srivastava. “Ninety-nine % of US companies are small companies, and plenty of are beneath stress from financial adjustments. They’re attempting to outlive, navigate uncertainty, and keep afloat.
On the similar time, shoppers are dealing with diminished entry to capital. “How will we assist each teams on this setting? That’s the place AI brokers are available — they’re constructed on platforms we’ve developed over a few years to unravel these precise issues.”
Srivastava used to run an artwork gallery in Mountain View and he mentioned that if he had entry to QuickBooks’ AI again then, with automated accounts receivable and payable, “I’d have completely used it.” With this background, he additionally stays optimistic about the usage of AI within the subject of artwork.
“It’s simply one other paintbrush.” Creativity has all the time discovered completely different retailers — music, artwork, literature, science, and maths.
“Some individuals might select to make use of it to specific themselves, others gained’t. I’ve shut associates who’re established artists, and so they see AI as only one extra medium, not a alternative. That’s how artwork has all the time advanced — new kinds emerge, individuals discover them, after which transfer on. That is no completely different.”
The publish Why Intuit Determined to Construct GenOS As an alternative of Onboarding 11,000 AI Fashions appeared first on Analytics India Journal.
