Moving AI from Pilot to P&L: What Finance Leaders Should Take Away From Expedition 2026

The month-end close has become the proving ground for AI in finance. Day 3 of Snowflake Expedition 2026 was built for the office of the CFO, with sessions featuring learnings from finance leaders who have moved AI into the work that defines their teams: closing the books, explaining variance, reforecasting, testing controls and preparing the tax provision.

Across every session, speakers made the same point in different ways. Finance will adopt AI when it can trust the output, and that trust comes from governed data, clear ownership and a human in the loop wherever judgment is required.

If you missed the live virtual event — or if you attended and want a little refresher — here are seven sessions that CFOs, chief accounting officers and finance executives will find useful as you start planning for next year. You can also check out the entire agenda track for finance leaders for more strategic and technical sessions on incorporating AI into your financial team workflows.

Finance teams are being asked to close faster, forecast with more confidence and advise the business in near real time, while also proving that AI investments pay off. These three sessions describe what that shift looks like and how leaders are managing it.

In the Day 3 opening keynote, Snowflake CFO Brian Robins shares his view of the trends and pressures shaping the finance function. Yasmeen Ahmad, Senior Vice President and GM of Data Cloud at Workday, introduces zero-copy integration between Workday Data Cloud and Snowflake, which makes Workday data directly available in Snowflake. Gordon Lee, Chief Accounting Officer at DoorDash, offers a Snowflake customer’s perspective on the impact of AI-driven finance workflows and how integrated data infrastructure, governance and permissions enables his team to use data more effectively. The keynote closes with a walkthrough of how Snowflake's own finance team runs on Snowflake, then follows a CFO through a continuous close cycle that runs from accelerating the close and explaining variance to reforecasting and preparing for earnings.

This session is a must-watch for any financial leadership team. It presents a shared, end-to-end picture of an AI-enabled close, which makes it much easier to agree on where your own team should start.

Watch “The AI Blueprint for Finance Teams” now.

Snowflake’s Brian Robins sits down with Steffan Tomlinson, CFO of Stripe, for a conversation about the changing role of finance and operations organizations. They discuss how to sequence capital allocation and measure ROI as AI moves from pilot to production, which human skills grow more valuable as automation takes on routine work, and how to move faster without giving up the accuracy and auditability the numbers demand. CFOs will appreciate hearing these tradeoffs from someone who has to defend them to a board, including the organizational and process changes that sit behind the technology.

Watch “The Finance Function, Reimagined: Leading Transformation in the Age of AI” now.

Brad Floering, SVP of Finance at Snowflake, lays out a practical approach to building an AI-first finance team. He covers how to choose and prioritize AI use cases and how to keep trust, governance and compliance intact as adoption spreads across the function. Share this session with your controllers and FP&A leads when you're ready to turn AI ambitions into a roadmap your team can execute.

Watch “Beyond the Pilot: A Finance Leader's Playbook for AI Transformation” now.

A fast answer has little value in finance if no one can trace, test and defend it. These sessions show how a global financial institution, an internal audit team and a Big Four tax practice are applying AI with accuracy treated as a hard requirement.

Gretchen Goers, SVP and Head of Finance Data and Automation at Northern Trust, explains how the firm is modernizing its finance function. She walks through the data challenges that prompted the change, the use cases her team prioritized first and how a trusted data foundation and AI are changing day-to-day finance work. For leaders in financial services and other regulated industries, this session offers a candid look at the wins, the lessons learned and what comes next for the office of the CFO at a global institution.

Watch “From Fragmented Data to Finance Intelligence: The Northern Trust Story” now.

Regulators, the board and SOX compliance all depend on numbers you can trace and defend, which puts internal audit in a difficult position when AI promises speed and scale. Amrita Kapoor, VP of Internal Audit at Snowflake, shares best practices for using AI in audit without compromising precision. She then describes how her team moved from sampling to continuous, AI-driven assurance with humans in the loop (for additional details, read our blog on how Snowflake’s internal audit team cut fieldwork by 50% with AI). Chief accounting officers and leaders who work closely with the audit committee will find a practical model for expanding coverage while keeping every conclusion explainable.

Watch “Rewriting the Audit Playbook: How to Scale Assurance with AI“ now.

Daren Campbell, EY Global Tax Data Leader and EY Americas Tax Technology and Transformation Leader, and Channing Flynn of EY’s Global Organization, Alliance and Ecosystems organization, explain how a unified tax data hub can create a single source of truth across the tax lifecycle. Their "transform once, reuse everywhere" approach supports compliance, provision, transfer pricing, forecasting and analytics from the same governed data. Tax team leaders will leave with a clear case for why lineage-enabled data has become a prerequisite for AI-enabled tax operations, informed by EY's work across many organizations.

Watch “Transformation at Scale: How to Build an Agentic Tax Team” now.

Every AI initiative carries a cost equation, and finance is increasingly the function asked to govern it. Maneesh Chhabra, Director of Value Advisory, and Pranav Pokharel, Principal Product Manager at Snowflake, share frameworks for managing AI spend down to the token level and for aligning FinOps discipline with generative AI adoption. CFOs and FP&A leaders will come away with a shared vocabulary and a set of controls that make it possible to scale AI across the business while keeping budgets predictable.

Watch “How to Manage Costs with Tokenomics and FinOps” now.

Once you’ve defined a strategy for incorporating AI into your finance workflows, it’s time to build the agents and tools that make it real. The finance-focused Builders track at Expedition showed the data engineers, analysts and architects who support finance how to get there, with demos built around real finance workflows and auditability treated as a design constraint. Here are a few sessions to forward to your FinOps and finance data partners.

• Build and Deploy Agentic Finance Applications with Snowflake CoCo shows how Snowflake's finance analytics team builds and ships agentic finance applications.
• Build and Automate Flux Analytics demonstrates how to automate variance analysis, one of the most time-consuming steps in the close.
• Build an Agent-Powered Tax System hows how Snowflake’s tax team built an agent to answer indirect tax, compliance and provision questions using CoCo, Cortex Search and Cortex Analyst. This session pairs well with EY's session on building an agentic tax team, and shows a governed tax agent at work.
• From Weeks to Hours: Automating Finance Data Pipelines with CoCo Skills on Snowflake covers automating the pipelines that feed reporting and planning.
• Get Your Data AI-ready: How to Access ERP Data with SAP Zero-copy Integration, presented by Hakkoda and TruQua (both IBM companies), explains how to work with SAP data in place rather than copying it using SAP BDC Connect to expose governed SAP data products in Snowflake.
• Close Faster, Audit Smarter: Build It with CoCo, Ask It with CoWork is a 60-minute virtual hands-on lab where builders learn how to land financial exports in Snowflake-managed Apache Iceberg tables, use CoCo to clean, standardize and match transactions across systems, and surface discrepancies on a Streamlit dashboard, then create a Semantic View and Cortex Agent so it’s accessible via natural language in Snowflake CoWork.

The finance leaders at Expedition 2026 advocated starting with a single workflow, usually the close or the forecast, proving the result on your own data and expanding from there. Choose the process that costs your team the most manual effort each month, watch these on-demand Expedition sessions with your controllers and data leaders, and talk with your Snowflake account team about scoping a pilot on your own data.