Levi's builds a 'super agent' to unify HR, finance, and IT

Key Takeaways

• Levi's is building a super agent that connects specialized AI tools across HR, finance, IT, and retail into a single employee interface
• Multi-agent workflows grew over 300% in recent months as companies moved from pilots to production deployments
• 43% of CFOs say agentic AI could reshape budget planning, with nearly half already using AI for cash flow monitoring

Levi Strauss & Co. is building what it calls a Super Agent to connect AI systems that currently operate in silos across HR, finance, IT, and retail operations. The goal: one interface for employees instead of four separate systems they must navigate manually. Multi-agent workflows grew more than 300% over recent months as organizations moved from pilots into production, according to Databricks data.

Why enterprise software stayed fragmented

Enterprise software organizes around functions. Finance has its system. HR has its own. IT runs on another stack entirely. Work that touches all three moves through each one separately, handed off by whoever knows which system to open next.

That handoff is where time and cost accumulate. An employee checking inventory, submitting an IT ticket, and initiating an HR process might touch three different applications with three different login flows and three different data models. The systems were never designed to talk to each other.

Levi's approach layers orchestration on top of existing infrastructure rather than replacing it. The company built specialized agents first, deploying AI tools across finance, design, and HR before adding the super agent as a coordination layer. According to a Microsoft customer story published June 4, employees will reach the right system through one entry point without navigating each separately.

“As a best-in-class direct-to-consumer retailer, the biggest thing that's changed for us is the speed at which we need to operate. This isn't just about a tool—it's a wholesale workplace transformation.”

— Jason Gowans, Chief Digital and Technology Officer, Levi Strauss & Co.

Single assistants vs. multi-agent systems

The distinction matters. Single AI assistants respond to prompts. Multi-agent systems manage workflows, passing tasks between specialized agents under defined rules. One produces an answer. The other produces an outcome.

Sheena Kunhiraman, Levi's vice president of HR technology and analytics, framed the value in practical terms: "Human/agent collaboration at Levi, I believe, is going to be all about augmentation—giving time back."

Goldman Sachs is applying similar logic in financial services. The bank is testing AI agents built with Anthropic's Claude to automate transaction reconciliation, trade accounting, client vetting, and onboarding. This work has resisted automation for decades because it requires processing large volumes of data against strict regulatory requirements.

How CFOs see agentic AI reshaping finance

The business case runs through finance as much as operations. PYMNTS Intelligence found that 43% of CFOs said agentic AI could have a high impact on dynamic budget planning. Nearly half already use AI to monitor working capital and cash flows.

The gap is between monitoring and acting. Agent networks can update projections, flag variances, and initiate adjustments within defined guardrails without routing each step through a human. That's the jump from dashboards to autonomous execution.

Ramp launched Applied AI Solutions in June for workflows that span multiple systems and require judgment when exceptions occur. Finance workflows are notoriously context-dependent. Every decision pulls from the policy, the vendor, the contract, the approval chain, and the exception history.

“In finance, every decision depends on buried layers of context: the policy, the vendor, the contract, the approval chain, and the exception history.”

— Ori Daniel, Head of AI Solutions, Ramp

Ramp's tool captures that context and turns it into agents that complete work within controls finance teams define. For companies already using automation tools like Zapier or Make to connect apps, this represents a step up: from rule-based automation to judgment-based execution.

What super agents can and can't do today

Super agents sit on top of existing infrastructure and coordinate across it rather than replace it. This is both the opportunity and the constraint. Companies don't need to rip out their ERP or HRIS. But they do need clean APIs and well-defined handoff rules.

The 300% growth in multi-agent deployments signals real production use, not just experimentation. But production brings new problems: agent governance, audit trails, error handling when one agent in a chain fails. These are early innings.

For finance teams specifically, the question isn't whether AI can read a spreadsheet. It's whether AI can navigate the exceptions, escalations, and judgment calls that make finance work hard. Ramp and Goldman Sachs are betting it can, at least within guardrails.

Frequently Asked Questions

What is an AI super agent in enterprise software?

A super agent is an orchestration layer that connects multiple specialized AI agents into a single interface. Instead of separate AI tools for HR, finance, and IT, employees interact with one agent that routes requests to the right system automatically.

How is Levi's using AI agents across departments?

Levi's built specialized AI agents for HR, finance, IT, and retail operations first. The company is now building a super agent to connect them, allowing employees to handle inventory checks, IT requests, and HR processes through one entry point.

What is the difference between AI assistants and multi-agent systems?

Single AI assistants respond to prompts and produce answers. Multi-agent systems manage entire workflows by passing tasks between specialized agents under defined rules. One gives information; the other completes work.

How are CFOs using agentic AI for finance?

43% of CFOs say agentic AI could reshape budget planning. Nearly half already use AI to monitor working capital and cash flows. Agent networks can update projections and flag variances without routing every decision through a human.

What challenges do enterprises face deploying AI super agents?

The main constraints are API readiness, agent governance, audit trails, and error handling when one agent in a chain fails. Super agents coordinate across existing systems rather than replace them, so clean integrations are essential.

Source: PYMNTS | / PYMNTS