Get faster salaries: How to help the new AI agents in Intuit companies get money up to 5 days faster and save 12 hours a month with an independent workflow

Intuit It has been on a trip over the past few years with AI Toulidi, and the merging of technology as part of its services into Quickbooks, Credit Karma, Turbotax and MailCheimp.

Today, the company takes the next step with a series of artificial intelligence agents that exceed this to transform the work of small and medium -sized companies in the market. These new agents work as a virtual team that automates the workflow and provides commercial visions in the actual time. It includes the capabilities of payments, accounts and financing that will directly affect commercial operations. According to Intuit, customers provide up to 12 hours per month, and on average, they will get a wage of up to five days faster thanks to the new agents.

“If you look at the path of our artificial intelligence experiences in Intuit in the first years, artificial intelligence is built in the background, however Intuit helpAshok Srivastava, the head of Amnesty International and Data Employee at Intuit, told Venturebeat: “I have seen a shift to provide information to the customer.” Now what you see is a full redesign. The agents are already working on behalf of the customer, with their permission. “

Technical Architectural Engineering: from beginners set to production agents

Intuit runs the path from auxiliary to AICERIC AI for some time.

In September 2024, the company Detail its plans To use artificial intelligence to automate complex tasks. It is an approach that was built firmly on the platform of the Obstetric IQ of the company (Genos), the basis for its efforts of artificial intelligence.

Earlier this month, Intuit announced a series of efforts that extend its capabilities. The company has developed its own Rapid improvement service This will improve information for any major language model (LLM). It has also developed what is called a smart perception layer for the institution’s data, which can understand the various data sources required for the functioning of the institution.

After a step forward, Intuit has developed a tool set for the start of an agent that relies on the technical basis of the company to enable AIC’s development.

Agent portfolio: from cash flow to customer management

With the presence of the technical basis, including the beginning of the agents, Intuit has built a series of new agents who help business owners to accomplish matters.

The Intuit agent’s wing explains the necessary technical development to move from an artificial intelligence to the implementation of the independent workflow. Each agent coordinates, NLP processing, and independent decisions in full commercial operations. Includes:

Payment agent: It works independently on the cash flow by predicting late payments, generating bills and implementing the follow -up sequence.

Accounting agentIntuit’s development of the rules -based systems to hold an independent notebook. The agent now independently deals to classify transactions and reconciliation and complete the workflow, and provides cleaner and more accurate books.

FinancingAutomation of strategic analysis traditionally requires dedicated business intelligence tools (BI) and human analysts. The main performance index analysis (KPI) provides and predicted the scenario based on how the company is established against peer standards while generating the recommendations of growth independently.

Intuit also builds customer axis agents who will help customer acquisition tasks. Treatment of salary statements as well as project management efforts is part of future issuance plans.

Beyond the user interface for the conversation: the design of the agent directed to the task

The new factors distinguish a development in how to introduce artificial intelligence to users.

The Intuit interface is revealed by the important user experience principles for publishing the institution’s agent. Instead of installing the possibilities of artificial intelligence on current programs, the company restructured mainly to the Quickbooks user experience mainly for AI.

“The user interface is now really directed about the work tasks to do,” explained Srivastava. “It allows visions and recommendations in the actual time to access the user directly.”

This approach centered on the task with the chat -based interfaces that dominate the AI ​​tools for the current institution. Instead of asking users to learn to demand demands or move in conversation flows, agents work in the current workflow. The system includes what is called “commercial summary” that gives the procedures and recommendations of the agent in the context.

Confidence and verification: the closed episode challenge

One of the most important aspects technically the implementation of INTuit addresses an important challenge in publishing an independent agent: verification and confidence. AI Enterprise Ai is often struggled with the black box problem – how do artificial intelligence agents perform properly when they work independently?

“In order to build confidence with artificial intelligence systems, we need to provide proof points to the customer that what they think is actually happening,” Srivastava stressed. “This closed episode is very important.”

The Intuit solution includes building the verification capabilities directly to Genos, allowing the system to provide evidence of the procedures of the agent and results. For the payment agent, this means that users clarify that the bills have been sent, track delivery and showing improvement in payment courses that result from the procedures of the agent.

This verification approach provides a template for the teams of institutions that publish independent factors in high risk commercial operations. Instead of asking users to trust in artificial intelligence outputs, the system provides auditable paths and measurable results.

What does this mean for institutions that are looking to access to Aiceric Ai

Intuit’s Evolution provides a tangible road map for the groups of institutions that plan for independent artificial intelligence applications:

Focus on completing the workflow, not the conversation: It targeted the operations for comprehensive automation instead of building chat facades for general purposes.

Building the infrastructure for the coordination of the agent: Invest in platforms that coordinate prediction, language processing and independent implementation within the unified workflow, not isolated tools of isolated intelligence.

Design verification systems in advance: Include comprehensive auditing paths and track results and user notifications as basic capabilities instead of subsequent incidents.

Workflow map before building technology: Use consulting programs for customers to determine the agent’s capabilities based on the actual operational challenges.

Plan to redesign the interface: Improving UX for the functioning of the agent instead of mobility patterns in traditional programs.

“When the large language models become a commodity, the experiments on which they are built becomes more important,” Srivastava said.