Get paid faster: How Intuit’s new AI agents help businesses get funds up to 5 days faster and save 12 hours a month with autonomous workflows


Join a reliable event by enterprise leaders in about two decades. VB Transform, Real Enterprise AI strategy brings together people who build. Learn more


Intuit Over the past few years, the technology is traveling in a few years by combining technology as part of their services in QuickBooks, Credit Karma, Turbotax and Mailchimp.

Today the company takes the next step with a number of AI agents to change how small and medium market enterprises work. These new agents work as a virtual team that automates workflows and provides real-time business concepts. These include opportunities for payments, accounts and finances that will directly affect work operations. According to Intuit, customers save up to 12 hours a month and will receive up to five days thanks to the average new agents.

“If you look at the trajectory of our experiments in the AI ​​Intuit in the first years, the AI ​​is built with the background and IntuitistA turn to inform the customer, “Ashok said that Srivastava, Intuit reported to Venturebeat at the senior AI and information officer.” What you see now is a complete redesign. Agents actually work with permission from the customer name. “

Technical Architecture: From the starting set to produced agents

Antuit assistants work on the road to the agent AI for a while.

Company in September 2024 details of their plans To use AI to automate complex tasks. This is a firmly built approach to the company’s generative AI operating system (Geno) platform, the foundation of the foundation of AI efforts.

At the beginning of this month, he announced a number of efforts to further expand the intuitic capabilities. The company has developed its own Quick Optimization Service This will optimize surveys for any major language model (LLM). The enterprise also developed a cleverly reporting cognitive layer for enterprise information that could understand various sources of information required for work flows.

A step forward has developed an agent built-in agent based on the company’s technical foundation to move forward, intuitively, intuitively.

Agent Portfolio: Customer management from cash flow

The agent was built by a series of new agents that helped businesses to do businesses, including the technical basis, including the initial sets of initial sets.

Intuit’s Agent Suite demonstrates the technical subtlety required to transform the execution of the autonomous work flow from the AI. Each agent coordinates the forecasting, natural language processing (NLP) and autonomous decision setting within full-time work processes. These include:

Payments Agent: Autonomous Autonomous Optimizes the flow of money by predicting the final payments, creating invoices and implementing follow-up sequences.

Accounting agency: According to the autonomous accounting, the rules represent the evolution of the intuitation from the systems-based systems. The agent now provides a mukhtronjada, reconciliation and completion of work flow, cleansing and more accurate books.

Finance: Traditionally automates strategic analysis that requires special business intelligence (BI) tools and human analysts. The company provides basic performance indicator (KPI) analysis, scenario planning and forecasting based on how the company works when making recommendations.

Intuit also builds customer HUB agents that will help customer shopping tasks. Salary processing, as well as project management efforts are also part of future free plans.

Out of spoken UI: task-oriented agent design

New agents record a evolution of how AI is presented to users.

The intuitive interface reveals the principles of an important user experience for the re-design, enterprise agent. Instead of tightening AI capabilities to existing software, the company is based on the QuickBooks user experience for the AI.

“The user interface is now really directed around business tasks,” said Srivastava. “It gives real-time concepts and recommendations to come directly to the user.”

This task-based approach is compared to the conversation-based interfaces that dominate current enterprise AI tools. Instead of learning referral strategies to users or touring negotiations, agents operate within existing workflows. The system includes the “working nutrition”, which called “work feed”, which contextifies the agent movements and recommendations.

Trust and Check: Closed-Loop Problem

One of the most technically important parties in the implementation of intuitiveness is a critical problem in placing an autonomous agent: inspection and confidence. Enterprise AI teams often fight the black box problem – Do you provide proper work when AI agents are autonomy?

“To build confidence with artificial intelligence systems, we must provide proof points to the customer what happened,” Srivastava said. “This indoor loop is very important.”

The solution of the intuitation is directly genos, which allows the system to provide evidence of agent actions and results. This means that it is demonstrated by the sending invoices to send invoices, improving the payment periods from the movements of the trailing and the agent.

This inspection approach offers a template for enterprise teams where high-rise agents placed in the work processes. If users are confronted by AI outputs, the system provides checkpoints and measurable results.

What does this mean for these businesses to enter the agent of AI

Intuit’s evolution offers a concrete roadmap for enterprise teams planning the autonomous EU applications:

Not talking, pay attention to the completion of the workflow: Target the target work processes for the final automation until the end of the completion of general-purpose chat interfaces.

Build an agent orchestral infrastructure: Investment, language processing and autonomous execution coordination, language processing and autonomous executive platforms within the unified workflows without isolated AI instruments.

Design verification systems from above: Extended audit tracks, the result follows and enter user notifications as more key opportunities later.

Map work before setting up technology: Use customer consulting programs to set agent opportunities based on topical operating problems.

Redesign for interface: Optimize UX for work flows managed by agent than traditional software navigation examples.

“As the great language models are recognized, the practices based on them are becoming more important,” Srivastava said.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *