Join our daily and weekly newsletter for the latest updates and exclusive content on industry-leading AI coverage. learn more
intuit – The financial software giant behind products such as Turbotax and QuickBooks is making great strides with Generative AI to enhance its offerings to small business customers.
In a technological landscape full of AI commitments, Intuit has built a proxy-based AI architecture that is delivering tangible business outcomes for small businesses. The company deploys so-called “done for you” experiences that handle the entire workflow autonomously and have a quantifiable business impact.
Intuit has been building its own AI layer called Generative AI Operating System (GENOS). The company details some ways to improve personalization using Gen AI VB Transform 2024. In September 2024, Intuit added Agent AI Workflow, this work improves the operations of the company and its users.
According to new Intuit data, QuickBooks online customers receive an average of five days of salary, with the full probability of overdue invoices being 10% higher. For small businesses whose cash flow is king, these are not just gradual improvements, but potential innovations that avoid corporate innovation.
Technology Trinity: How Intuit’s Data Architecture Enables Real Proxy AI
What distinguishes the Intuit approach from its competitors is its sophisticated data architecture designed specifically to enable an agent-based AI experience.
The company built what CDO Ashok Srivastava calls the “Trinity” of data systems:
- Data Lake: The basic repository of all data.
- Customer Data Cloud (CDC): A service layer dedicated to artificial intelligence experience.
- “Activity Bus”: The streaming media data system enables real-time operations.
“The CDC provides a service layer for the AI experience, and then the data lake is the repository of all this kind of data,” Srivastava told VentureBeat. “The agent will interact with the data, and it has a set of data that can be viewed to attract information.”
Beyond vector embedding into power proxy AI
The Intuit architecture is diverged from typical vector database approaches, and many companies are rushing to implement it. While vector databases and embeddings are important for powering AI models, Intuit recognizes that true semantic understanding requires a more comprehensive approach.
“Where the key issues continue to exist is basically ensuring we have a good, logical and semantic understanding of the data,” Srivastava said.
To achieve this semantic understanding, Intuit builds a semantic data layer on top of its core data infrastructure. The semantic data layer helps to provide data and meaning beyond the data itself or its vector representation. It allows Intuit’s AI proxy to better understand the relationships and connections between different data sources and elements.
By building this semantic data layer, Intuit is able to enhance the functionality of its vector-based system by having a deeper, deeper contextual understanding of the data. This allows AI agents to make smarter and meaningful decisions for customers.
In addition to basic automation: How agent AI automates the entire business process
Unlike businesses that implement AI for basic workflow automation or customer service chatbots, Intuit focuses on creating a fully agented “done for you” experience. These applications can handle complex multi-step tasks while requiring only the ultimate human approval.
For QuickBooks users, the proxy system analyzes customer payment history and invoice status to automatically draft personalized reminder messages, allowing business owners to simply review and approve them before sending. The system’s ability to personalize based on relationship context and payment methods directly leads to faster payments.
Intuit is applying the same agency principle internally to develop independent procurement systems and human resources assistants.
“We have the ability to carry out an in-house agent procurement process that can be used to buy supplies and book travel,” Srivastava explained.
Design for the era of inference model
What is possible to give Intuit a competitive advantage over other enterprise AI implementations is how the system’s design emerges with vision for advanced inference models such as DeepSeek.
“We built Gen Runtime and look forward to the inference model coming soon,” Ashok revealed. “We are not behind eight goals…we are ahead. We build the ability to reason assuming that there is a presence.”
This forward-looking design means Intuit can quickly incorporate new reasoning capabilities into its agent experience at the time of its emergence without the need for a building overhaul. According to Srivastava, Intuit’s engineering team is already using these features to enable agents to reason about large amounts of tools and data in ways that were previously impossible.
Shift from AI hype to business impact
Perhaps most importantly, Intuit’s approach explicitly focuses on business outcomes rather than technical performance skills.
“There is a lot of work in AI itself today and a lot of fanfare that has completely changed the world, and all of which will completely change the world, and I think that’s good,” Srivastava said. “But I think it’s better to show that it’s actually helping the real people do better.”
The company believes that deeper reasoning capabilities will enable a more comprehensive “done for you” experience that covers deeper customer needs. Each experience combines multiple atomic experiences or discrete operations to create a complete workflow solution.
What does this mean for businesses that use AI
For businesses looking to implement AI effectively, Intuit’s approach provides several valuable courses for businesses:
- Focus on the results of technology: Instead of demonstrating AI with measurable improvement goals as target specific business pain points.
- Keep the future model in mind: Design architectures can be combined with emerging inference functions without complete reconstruction.
- Solve data challenges first: Before you rush to implement a proxy, make sure your data base supports semantic understanding and cross-system reasoning.
- Create a complete experience: Go beyond simple automation to create an end-to-end “done for you” workflow that provides a complete solution.
As agent AI continues to mature, by focusing on complete solutions rather than isolated AI capabilities, businesses following Intuit’s role model may find themselves achieving similar specific business results, rather than simply creating a technology buzz.
Source link
