Intuit will show off how it rebuilt its AI infrastructure to support fast and complex tasks at VB Transform 2026

2026-06-24 · Source: VentureBeat · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Cloud Computing & IT Infrastructure · Depth: Advanced, quick

Summary

Intuit has significantly overhauled its AI infrastructure to support complex agentic AI tasks, moving away from a multi-agent system that prioritized broad capabilities. This strategic shift involved adopting a granular, skill-and-tool-based architecture, crucially embedding human experts directly into the AI workflow. Nhung Ho, Intuit's VP of AI, will detail these technology decisions, including the creation of an abstraction layer and the decoupling of orchestration from specific model providers, at VB Transform 2026 on July 14 and 15. This new architecture enhances agility, enabling Intuit to integrate optimal tools from various sources. The conference will also feature sessions from Target, Instacart, Asana, Rivian, Databricks, and Atlassian, highlighting broader industry focus on agentic AI orchestration.

Key takeaway

For AI Architects designing next-generation systems, Intuit's shift demonstrates the necessity of moving beyond monolithic multi-agent setups. You should prioritize granular, skill-based architectures that integrate human expertise and decouple orchestration from specific model providers. This approach enhances agility, allows for optimal tool selection, and ensures your infrastructure can scale to meet complex agentic AI demands, avoiding the limitations of legacy systems.

Key insights

Intuit's AI infrastructure overhaul prioritizes granular, skill-based agents with human integration, enabling complex task handling and provider independence.

Principles

• Legacy architectures hinder complex agentic AI.
• Decompose large agents into specialized components.
• Decouple orchestration for agility and tool flexibility.

Method

Intuit's method involved decomposing multi-agents into granular, skill-and-tool-based components, integrating human experts, and rebuilding the orchestrator, planner, and "brain" to support complex agentic workflows and provider-agnostic orchestration.

In practice

• Integrate human experts into AI workflows.
• Build abstraction layers for model provider independence.
• Select optimal tools from diverse providers.

Topics

• Agentic AI
• AI Infrastructure
• AI Orchestration
• Multi-agent Systems
• Human-in-the-loop AI
• Model Agnostic Architecture

Best for: CTO, VP of Engineering/Data, AI Architect, Director of AI/ML, MLOps Engineer

Related on AIssential

• WWDC 2026 and the Shift from AI Models to Intelligent Systems — The AI industry is shifting from model-centric development to building integrated intelligent systems for seamless user experiences.
• Multi-Agent AI Is Having Its Microservices Moment — Multi-agent AI systems offer modularity and scalability, akin to microservices, for complex AI applications.
• The Agentic Shift: Why Engineers Are Becoming Workflow Architects — Agentic AI transforms engineering by autonomously planning and executing tasks, requiring human oversight for safety.
• The “Computer” Clash: Perplexity vs DevRev on True AI Autonomy — True AI autonomy requires deep data integration and the ability to take action, not just provide answers.
• [AINews] New AI Infra unicorns: Exa, Modal, TurboPuffer — The AI ecosystem is rapidly maturing, marked by significant infrastructure funding, model advancements, and agent-centric product development.
• The AI Labor Shift: Why We Are Transitioning from ‘Doers’ to ‘Reviewers’ — Generative AI shifts labor from task execution to human review and system orchestration, demanding contextual judgment.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by VentureBeat.