How AI is supercharging finance teams: Best practices on adoption, hiring, and risk mitigation

Change happens gradually, then suddenly, to use the words of Ramp CEO Eric Glyman, and that rapid shift is now underway in finance.

At OnRamp New York 2025, finance leaders showed AI is already driving productivity gains and shared best practices to help their peers push AI initiatives from experimentation to impact.

3 strategies to embed AI in finance

While finance may still carry a reputation for being risk-averse, the reality is these teams are at the forefront of AI adoption at many companies. They’re using AI to automatically answer common vendor queries, create financial reports, and code transactions. Here are proven approaches to get these initiatives off the ground:

• Allocate time for experimentation: At wedding planning website Zola, the finance team stole an idea from its engineering team with Guild Days — dedicated time to explore and test how AI and other technologies could make time-intensive processes more efficient.
  
  “Finance is one of those areas where you always have deadlines, such as month-end close and reporting to send to your board,” Zola VP and Controller Joe Horn said. “So to carve out two days and just solely focus on these AI and automation projects in a team setting was really good.”

• Explore new AI features in existing tools: Software is evolving quickly with AI-powered capabilities, and businesses have much to gain by keeping up with these releases and being among the first to test them. Datadog, which provides application monitoring software, has found ways to empower its team with the AI capabilities available in its existing ERP, CRM, and close management systems.
  
• Invest in learning and development: Go-to-market automation platform Clay issued employees Ramp cards to explore tools and enroll in classes that could level up their AI skills. Participation wasn’t mandatory, but most employees took advantage — it was an easy way to get comfortable with tools and find more use cases for AI.
  
  “It gives them a sandbox within which they can play, and if they ever see a tool that they need to elevate to the rest of the team, they can come and request specific budget and pull that up to the broader organization,” Clay Head of Finance Karan Parekh said.

Set realistic expectations and keep iterating

Successful AI rollouts start with the right expectations. That means realizing that new AI-driven tools and processes won’t always bring immediate success and not giving up on these tools after a few underwhelming experiences.

“I think many people use AI and treat it like it's a nepotism hire that they had to make,” said Ravi Gupta, a partner at Sequoia Capital. “You give them a mediocre piece of work that doesn't mean anything and the first time there's an error, you're like, ‘Ah, I knew it! I knew they couldn’t handle it! I'm never going to give them anything of substance again!’”

Gupta recommended you instead think about AI as a super-talented hire you’ve been pursuing for years. If that person made a mistake, you would give them the benefit of the doubt rather than immediately questioning their skills. What important context or information did you not provide that led to the mistake?

Hiring for finance in the AI era

Throughout the hiring process, finance leaders are looking for both AI literacy and the critical thinking skills that AI cannot replace. They’re encouraging candidates to use AI tools for take-home assignments or interview prep, though they also want to see that they added their own touch — that there aren’t big gaps in explanations or obvious errors. They use interviews to assess critical thinking skills by digging into how they thought through the assignment and potential challenges their plan could present.

“You need to have the right grounding and frameworks to think through a problem critically,” Parekh said. “And then AI is an enabler to get to the facts that you need so you get to the right decision in a more expedient way.”

Put simply, speakers agreed that AI is an accelerant for individuals who already have strong fundamentals.

“If someone who is a good, productive finance employee knows how to utilize AI, they'll be that much better,” Horn said.

Addressing common AI-related risks

The potential risks AI presents has deterred some businesses from leaning in with AI-forward finance teams. The key is to find a balance between quickly empowering teams and maintaining guardrails, which will look different for your business than another company.

Many of the concerns around AI in finance come back to compliance. To reduce data security risks Datadog, a large, publicly traded company, has its security team review any technology that touches core systems holding customer data. Zola, while private, relies on engineering leadership to set controls on and monitor Google Gemini use.

Another piece of compliance is audits, and panelists noted auditors are now asking about AI use and how they’re verifying its outputs. Think about whether you would have solid answers to these questions and, if not, what changes you need to make.

Since finance is ultimately responsible for cost controls, they should closely monitor AI tooling costs that can climb quickly and understand why the business is spending where it is. Clay had had some unexpectedly high LLM bills when teams used a more advanced — and more expensive — model than necessary for the job, Parekh noted.

Throughout the day, leaders showed that the promise of AI in finance is real. It can keep teams lean and hyper-productive while shifting their focus from operational to strategic work. But realizing this potential takes thoughtful coordination and planning — along with an understanding that AI is not a magical cure-all.

For even more insights on how finance can leverage AI, watch the full customer panel with Zola, Clay, and U.S. Soccer.