Shopify’s finance team goes from reflexive AI usage to being an AI-powered organization

Jeff Hoffmeister, CFO at Shopify, recently sat down with Zip’s CEO Rujul Zaparde at Zip Forward 2026 to discuss what it takes to run a finance organization in the age of AI.

The challenge

Jeff Hoffmeister, who spent 22-23 years at Morgan Stanley before joining Shopify, said the first thing that struck him was the pace of decision-making. As Jeff shares, "for those of you that work at founder organizations, that pace of decision-making is just different than a big corporate America organization."

That pace set the tone for a bigger shift. Just over a year before the event, Shopify’s CEO circulated a memo on how the company thinks about AI. Within that memo, there was a mandate that no one could hire additional employees until they prove AI can't do the job. 

Additionally, Jeff said, the memo noted everyone must use AI, whether they work in finance, legal, HR or sales. The expectation was immediate: "all people, everywhere in the organization, use it reflexively."

For finance, the mandate raised several questions:

• ‍Moving beyond "just using it". Shopify's performance review approach looks at whether people use AI, whether they use it effectively, and whether they turn that use into real impact.‍
• Looking forward, not only back. Jeff wanted to look at things beyond month-end and quarter-end results. Jeff wanted AI to also improve FP&A work like projections, forecasting and pricing.‍
• Trust. Many finance leaders don't fully trust AI on accounting, where tolerances matter. Without trusted sources, Jeff said, "I could ask three people the same question and get three different answers."‍
• Finding the time. Nearly every leader says they want to be more proficient in AI, but as Jeff notes, "where do you carve out the time? That's the hard part."

The solution

Hoffmeister's playbook starts with people. About a dozen mostly mid-level finance employees leaned in first, including an accounts payable clerk who built a replica of a long-established enterprise application in about 18 hours. Every Friday the team meets for an hour to demo what they've built. "Probably every other week someone comes and says, 'Oh, we used to have this process that took two days or two weeks, and now it takes three minutes.'"

Next comes the data: trusted sources, tied to the right MCP servers. A couple of months in, employees go from "assuming the answer they get from AI is wrong, to assuming the answer is right," with humans still checking financial statements.

Finally, he asked his top 50 leaders to name the most toilsome part of their job and fix it with AI. About 85% "were able to do something of real significance."

Visibility helps too. Shopify's internal AI assistant, River, fields questions in a shared finance Slack channel, so everyone sees what colleagues ask and gets ideas of their own. The team also narrowed its tools to two or three platforms so people can go deep on one.

Shopify has used Zip since 2023, from intake and procurement orchestration through procure-to-pay. For Hoffmeister, the value is context. He sees the full approval chain, can tag colleagues in Slack, and the next approver sees every earlier comment, so he can ask "is this an approval for a number, or can I figure out all the context I need?"

It also raises the questions finance can ask: who sends up weak approvals, whether an agency has run over budget across three years of bids, how software rates benchmark. "We are in a position now to ask much more sophisticated questions and do much more analysis," he said, calling it "completely up-leveling."

The results

The payoff shows up in speed and in the quality of the conversation. Analysis that once took two weeks of slide-building now comes together in five minutes, and finance, data and sales work together on questions like salesperson impact and the length of sales leads. Those analyses have changed how Shopify thinks about its sales force organization, customer profitability in a geography and cross-selling. "You actually feel like you're moving the company forward in a way that people didn't feel 18 months or two years ago," Jeff says.

Zip supplies the context behind those decisions. Approvers can see the full chain of comments and questions rather than just a number, and Hoffmeister points to procurement questions, like how often an agency ran over budget across three years of bids, as the kind AI can now answer quickly with that context. 

Looking back on Shopify’s finance transformation, Jeff offers the following advice to other finance leaders: "lead the team there, live in that future, and make it happen today."