Intuit scrapped its own AI agent architecture twice in four months. At VB Transform 2026, its AI VP called that the fast path
Our take
Intuit’s recent architectural pivots in their agentic AI journey, as detailed at VB Transform 2026, offer a fascinating and frankly, reassuring glimpse into the realities of building complex AI systems at scale. The company was an early pioneer in embracing this approach, but their experience underscores a crucial point: the path to effective agentic AI isn't a linear progression of innovation; it’s a process of iterative refinement and, sometimes, significant course correction. This journey mirrors the broader challenges facing organizations exploring agentic AI, particularly the need to move beyond theoretical frameworks and grapple with practical limitations. It's particularly relevant given the increasing attention being paid to agentic AI frameworks like OpenClaw, as highlighted in Brex built its AI agent policy by watching what agents actually do, not by writing rules first, showcasing the importance of data-driven policy development.
The core takeaway from Intuit’s experience is the inherent fragility of centralized orchestration layers. The compounding error problem, where each handoff between agents in natural language leads to a degradation of context, is a surprisingly elegant explanation for a system's failure. It’s a reminder that even well-intentioned architectural choices can introduce unexpected bottlenecks and vulnerabilities. The shift to a skills and tools-based architecture, while requiring a significant engineering effort – a 60-day rebuild – demonstrates a pragmatic response to this challenge. Importantly, Intuit’s leadership successfully navigated the internal resistance to scrapping recent work, leveraging compelling demonstrations based on real customer data to justify the change. This highlights the pivotal role of aligning technical decisions with tangible customer outcomes—a principle that should resonate deeply with any organization pursuing AI-driven transformation. Their approach stands in contrast to some of the more hyped promises surrounding “revolutionary” AI solutions, emphasizing instead a focus on practical problem-solving.
What’s particularly compelling is Intuit’s innovative approach to feedback. Moving from sparse, often negative feedback to a continuous stream of data from every customer conversation represents a paradigm shift. This granular level of insight, coupled with the development of models to analyze that feedback, promises a much more iterative and responsive development cycle. The willingness to confront the blunt honesty of customer feedback – “They straight up tell you, ‘You suck. I hate this. This is not right’" – is a testament to a culture that prioritizes continuous improvement. Furthermore, the integration of human agents into the conversation flow, allowing customers to seamlessly connect with support personnel or accountants, directly addresses a common caveat in AI chat interactions, turning a potential disclaimer into an opportunity for enhanced service, as evidenced by similar efforts in other areas like Roblox launches an AI-powered game-creation feature in its mobile app.
Ultimately, Intuit’s journey underscores the iterative and often circuitous nature of AI development. The repeated rebuilds, the internal debates, and the reliance on customer feedback are all hallmarks of a realistic approach to building transformative technology. As agentic AI continues to evolve, it will be crucial to prioritize adaptability and a willingness to abandon flawed assumptions, even if it means discarding significant prior investment. The question now is, how quickly will other organizations learn from Intuit’s experience and embrace a more pragmatic, feedback-driven approach to agentic AI development, avoiding the pitfalls of overly ambitious architectural designs?
Intuit was an early pioneer in the usage of agentic AI, but its path to success has hardly been a straight line.
At VB Transform 2026, Intuit VP of AI Nhung Ho described how the company rebuilt its agent architecture twice in the span of about four months, first moving from a fleet of specialist agents to a central orchestration layer, then abandoning that layer for a skills and tools based system once the orchestrator itself started failing under its own complexity. The full second rebuild took 60 days, with a first working version in under 20.
The failure mode that forced the second rewrite was specific. Agents in the orchestrated system passed results to each other in natural language, and each handoff lost context the next agent needed to act correctly.
"If you have 10 agents and they all are passing to each other, every time that pass happens, error compounds," Ho said.
Why the orchestration layer broke down
Ho said the original push toward specialist agents came from a straightforward customer complaint. A fleet of capable agents is still something a customer has to manage, deciding which agent to use for which task. Intuit's answer was a system that could take a task and route it internally, without asking the customer to pick an agent themselves.
That orchestration layer held up for about three months, which Ho described only half joking as roughly a year in the compressed timeline of agent development in 2026.
It broke for a structural reason rather than a capacity one. Passing outcomes between agents in natural language meant each downstream agent had to infer how the upstream agent reached its conclusion, and that inference degraded with each additional hop. A ten agent chain did not fail occasionally, it compounded errors by design.
That diagnosis is what sent Intuit back to a skills and tools architecture.
The 60-day rebuild, and what it took to get engineering buy-in
Rebuilding a production agent system in 60 days required more than an architectural decision. Ho said the harder problem was internal, convincing both leadership and the engineers who had built the original agents that scrapping recent work was the right call.
The pitch to leadership relied on evidence rather than argument. Ho's team built a demo of the new architecture using real customer queries pulled from production, then showed it performing better than the existing system on the same tasks.
"The best proof, at least my belief, is what are customers trying to do? And whatever system you build needs to address those problems," Ho said.
Winning over engineering required a different case. Hundreds of engineers outside Ho's core team had built the specialist agents being retired, and the ask was to take their agents apart into individual skills and tools instead.
Ho said the motivating argument was scale. A standalone agent solved one narrow problem, while a shared skill or tool built into the new architecture could serve every customer who touched that part of the product. That shift also changed what partner teams were responsible for day to day, moving their focus from building agents to running evals, since evals became the only way to measure whether the new architecture was actually working.
Bringing a human into the loop, and feedback at a different scale
The clearest customer facing result of the rebuild is a feature that lets a live agent conversation pull in a human — though it's currently in early testing, live to about 1% of Intuit's customer base. "We're going to be scaling it up in the next few weeks," she said.
Ho said a customer can bring in an Intuit product support person mid conversation, or their own accountant, or one of Intuit's own bookkeepers, and that person joins with the full context of what the agent has already done.
Ho drew a direct contrast with how most AI chat products handle the same situation. A general purpose assistant answering a tax question typically ends with a disclaimer to consult a professional. Intuit's system is built to connect the customer to that professional directly, inside the same conversation.
That human handoff sits alongside a permissions model built for financial data specifically. Every action an agent takes on a customer's financial data requires explicit permission first, though Ho said that requirement can ease over time as customers build trust in the system. Intuit keeps an audit log of everything an agent does that can be reversed if needed.
Feedback in the agentic AI era
The rebuild also changed how Intuit gathers and uses feedback, a shift Ho said is qualitatively different from what came before.
"Feedback in the past used to be very, very sparse, and it was also very bimodal," Ho said. "Either they loved it or they hated it, and usually it tends towards the negative."
In a chat based system, every conversation functions as feedback, which Ho said moved the company from roughly 0.3% of customers ever giving explicit feedback to something close to 100%.
Ho said she has returned to writing code herself specifically to build models that analyze that feedback volume systematically, looking for where the system is falling short at a scale no manual review process could keep up with.
That volume comes with a tone most product teams aren't used to hearing directly. Customers tell the agent exactly where it failed, in plain terms.
"They straight up tell you, 'You suck. I hate this. This is not right,'" Ho said. "But they're also willing to give the systems grace and correct it as well, and so the onus is on all of us to harvest this new piece of feedback and type of feedback, and actually improve the system."
Read on the original site
Open the publisher's page for the full experience