Benchmark your company's expenses with Ramp's data.
straight to your inbox
Annual planning is in full swing—and CFOs everywhere are getting asked how AI should fit into their budget.
How much should companies spend on AI to accelerate their business operations next year? What use cases will drive value, and how can they be responsibly vetted? How can leaders drive internal adoption, so new tools actually get used?
These questions are likely coming from all sides—your CEO, the board, and employees. To dig into this topic, I recently hosted a fireside chat with Johnnie Walker, Co-founder & Director at Rooled, and Brian Gwiazdowski, VP of Portfolio Strategic Finance at Primary Venture Partners. They brought a range of perspectives to the table, with Johnnie’s fractional CFO expertise and Brian’s experience on the investor side. Together we discussed five ways for finance leaders to steer these crucial conversations.
1. Use controlled experimentation to ensure value-added AI investments
With new AI tools emerging daily, many CEOs and boards are pushing their organizations to invest more in operational AI. In this exciting environment, CFOs are responsible for implementing fiscal guardrails.
One way to do so is to set a specific budget for AI experimentation. This constrains potential risk to your budget while allowing uncapped opportunity for innovative use cases. For example, organizations could give employees $50 monthly to spend on AI tools. Hint: use Ramp cards to easily control employees’ access to this budget and ensure proper usage.
Brian emphasizes the need to track outcomes carefully. “You want to make sure teams aren’t buying tools that they don’t end up using and that you’re getting results from those you are,” he says. Johnnie adds that Ethan Mollick’s concept of AI co-intelligence is a useful starting point. Instead of pushing your team to replace work with AI, “in everything you do, you can be including AI,” he says. “Once you get that bubbling up of interest, along with top-down recognition that other companies are doing this, you'll find that things start to grow.”
Experiment—inspiration for amazing ideas will come from daily practice.
2. Create guidelines and policies to drive responsible AI
Beyond fiscal guidance, finance leaders should also play an active role in steering their organizations toward sound, responsible use of AI. That starts with crafting guidelines and policies in collaboration with their legal and CTO counterparts.
As a first step, create a policy for consumer-grade AI assistants like ChatGPT. Ramp data show spending on these tools has grown 6x year over year, signaling these tools are gaining a foothold in the workplace. You may want to specify that employees can use these tools for first drafts and iterations—not final copy. Your policy should also warn against using sensitive financial or customer data for prompt engineering.
Brian wisely notes, “This is a fast-moving space, so it’s not possible to just open your laptop and independently write the perfect AI policy. The best approach is to talk to peers and experts, and do your best to output a sensible policy.” To that end, Rightworks offers a great template, covering everything from allowable tools and use cases to data guidelines and output review.
3. Set clear AI procurement rules to influence vendor selection
CFOs should also take a strong stance on how their teams procure AI vendors. Right now, hot new AI vendors are a dime a dozen. Employees need help to discern which products are truly innovating—and therefore a good use of budget—and which are simply using AI as a buzzword to drive hype.
When reviewing new vendor requests, pay attention to vendors’ product roadmap, velocity of innovation, and ability to integrate new technology. Do they have a plan to keep evolving, or will you be stuck with a product built on outdated technology from five years ago?
Johnnie shares additional advice: “If a product says it’s using AI, see if they can explain the technicalities, how the data flows. Any finance tool is going to be licensing infrastructure from big vendors like OpenAI or Anthropic. It’s entirely justifiable to ask how they’re using it, and what makes them different from other vendors.”
4. Get employee buy-in with transparency and education
Your CEO and board may not need much convincing about AI—but your employees may be a tougher sell.
Earlier this year, a survey by Retool found just 27% of entry-level employees want to invest more in AI, likely due to worries about their existing jobs and tasks being made redundant. To temper these worries, leaders should emphasize humans’ and machines’ complementary—not competing—strengths. Gartner has a great visual for this human-machine learning loop that showcases the strengths of both types of workers:
In many ways, AI isn’t any different from other types of digital transformation that CFOs have had to shepherd in recent years. Employees should be actively included in the adoption conversation and given a holistic understanding of how and why these tools will be rolled out at their organization. “You want teams to be taking part in any digital transformation, including AI,” says Brian. “Upskilling employees with AI, and empowering them to experiment, is a great way to get that buy-in and make sure you’re ready for whatever future comes along.”
5. Demonstrate leadership with a clear finance AI roadmap
Historically, finance has lagged behind other functions in AI adoption and experimentation. But that represents a golden opportunity for CFOs to lead the way and brainstorm exciting AI use cases.
Johnnie, Brian, and I all agree there’s no magic app that will close your books for you just yet. But there are a flurry of AI applications that can help your team decrease time to close, strengthen accuracy, and increase the efficiency of your current team.
- Start by automating the transaction layer. Naturally, Ramp is a great option on the expense side. Our AI matches transactions to contracts and bills, reviews them for compliance, and automates account coding. On the revenue side, look into tools like Tabs for revenue recognition automation and Stripe Radar to fraud prevention.
- Auditing is another powerful application. “AI can look at every single transaction, as opposed to traditional internal or external auditing, which is a sample-based approach,” says Gwiazdowski. Trullion is one exciting name to keep an eye on in this space.
- AI is also well-suited for variance analysis within FP&A. “AI could do the first version of a BVA or financial report, which Finance can edit before sending to investors or internal stakeholders,” says Walker. You could also apply the same approach to writing investor reports and MD&As.
- If Finance oversees Legal, contract review is a prime opportunity to save time and money with automation. Tools like Ironclad or wordsmith.ai can help review contracts in seconds to identify non-permissible risks or clauses. Some even go the next step by suggesting redlined language.
AI hasn’t reinvented the finance function just yet. But new solutions don’t need to be revolutionary to offer value. Ultimately, boards are looking for operating leverage to positively impact operating margin. Johnnie says it best: “Our customers want good financials, they want financial insights—they're not necessarily asking us to drive that with AI. But if we can use AI to deliver that, that drives efficiency for our firm.”
Watch our full conversation
It’s an exciting time to be a CFO. AI is paving the way for finance leaders to create better financial outcomes through newfound efficiency and innovation. Check out the fireside chat for even more insights.