See how real CFOs are using Microsoft Copilot, ChatGPT (OpenAI), and Gemini to build and iterate on formulas to create a more consistent and efficient close process.
|
April 5, 2026
.webp)
Finance teams have found a way to close the books faster. The answer is AI.
That's one of our big takeaways from one-on-one conversations we had with 30+ CFOs for SpendHound's 2026 CFO AI Report: Finance teams are using AI to automate formulas, stitch together data, and maintain spreadsheet logic - cutting tasks that used to take hours down to just minutes. The result is a close process that's faster, more repeatable, and less dependent on any one person's spreadsheet knowledge. And with the time saved, teams can spend more time on valuable tasks like financial analysis of the data.
Here's how the CFOs we spoke with are using AI to speed up the close process.
For this use case, we observed how CFOs use general-purpose LLM tools to simplify and streamline close processes. General-purpose LLMs (large language models) are AI tools that can generate text, formulas, and code from natural-language prompts, making them useful across a wide range of finance workflows.
These tools don’t replace ERPs or more traditional finance systems. Instead, they sit alongside them, helping to automate the most manual work: generating formulas, writing scripts, and debugging workflows.
Based on SpendHound’s SaaS pricing benchmarks — built on actual contract data from 1,000+ companies — ChatGPT (OpenAI) costs SMB teams (50–999 employees) around $42,000 annually and enterprise teams approximately $167,000. Microsoft Copilot is typically bundled into Microsoft 365 and Gemini is typically bundled with Google Workspace, making these two general-purpose LLMs effectively zero incremental cost for teams already on those platforms.
(Note: These numbers were accurate as of the time of the publication of the full report. For the most up-to-date real-time prices for ChatGPT (OpenAI) and other AI and SaaS software tools, check out our Marketplace.)
.png)
According to a recent report by CFO.com, 50% of finance teams still take over a week to close the books, largely as a result of issues with data, technology use, spreadsheets, and cash reconciliation.
The CFOs we spoke with agreed, putting special emphasis on the fragmentation of underlying data.
The inputs required to close the books live across multiple systems, each with its own structure and logic. Finance teams end up acting as the connective layer, pulling data from ERPs, billing platforms, and operational tools, then formatting it into something usable.
That creates a process that technically works, but carries a lot of hidden overhead. As businesses evolve — with new products, new revenue streams, changing reporting needs — the process gets less and less efficient, with much time spent preparing and stitching together data and not enough time analyzing it.
That underlying complexity shows up most clearly in how close workflows are actually executed.
For most teams we spoke with, the close is still driven by spreadsheet-based models that handle revenue recognition, deferred revenue, and internal reporting. These models contain layers of logic built over time, connecting data across systems through manually constructed formulas.
Where things start to break down is in iteration. Even small changes — like a new reporting view or updated assumption — can require finance teams to revisit core parts of the model. That might mean rewriting formulas, restructuring tabs, or rebuilding sections entirely.
That ongoing effort is what slows down the close and introduces delays.
Instead of building logic from the ground up, some of the CFOs we spoke with describe how their finance teams are starting with AI-generated formulas and scripts. You start by describing the outcome you need and the tool produces a working version that can be refined. This means your starting point is a solid working draft that can be tested and adjusted instead of a blank sheet. In a recent in-person Vibe Coding event we hosted, we saw this in real time: Standing up a visual representation of a spreadsheet makes it easier for CFOs to catch errors and ideate.
As one CFO explained,
“Rather than trying to build a formula myself, I tell the tool what I need — and within seconds, I have the right formula.”
AI is also helping with debugging and iteration. When something breaks or needs to be modified, teams can resolve issues through prompting instead of trial-and-error. That reduces the friction that typically slows down the close.
The finance leaders we spoke with are methodical about ensuring that the addition of AI is resulting in real measurable gains.
They define ROI in practical terms: shorter close timelines without adding headcount, and less reliance on manual intervention to get through the process.
And those outcomes are starting to show up. CFOs reported that their teams are spending less time writing and maintaining formulas, and less time reworking models when assumptions change. Tasks that previously took hours — especially those involving complex joins or revenue scenarios — now wrap in minutes. Over time, that reduction in manual effort translates into faster, more predictable close cycles.
That’s why finance teams are also validating results against the underlying data. They look for outputs that reconcile back to source systems, particularly the ERP, and for workflows that produce consistent results across reporting periods. Instead of relying on last-minute fixes or deep familiarity with a specific spreadsheet, teams are able to run processes that behave the same way each period.
Faster execution plus consistent, reliable outputs equals clear ROI.
Before introducing AI into your close workflow, it’s worth stepping back and making sure you’re solving the right problems — and that your team will experience real efficiency gains from the effort.
Here’s how to prepare:
Even when finance teams want more real-time visibility, the effort required to pull the data mid-cycle often isn’t worth it, so everything gets deferred to the close.
One CFO we spoke with described it this way: “If during the month I want to look at costs, the difficulty is that these costs sit in myriad places. AWS in one place, ClickHouse in another, BILL in another. That’s not the best use of my time in the middle of the month — so I just wait for accounting to close.”
AI starts to change that.
When it becomes easier to build, test, and adjust financial logic, the cost of answering questions drops. It speeds up the close and expands what’s possible between closes.
For CFOs evaluating where AI fits, this is one of the clearest examples: take a process that happens every month, and make it meaningfully easier to run.
Download the full report for more practical applications of AI: 2026 CFO AI Report: 7 Proven AI Use Cases from 30+ CFOs.
See how SpendHound can help you pay the right price for ChatGPT, Copilot, and other tools to support close workflows.
How are finance teams using AI to reduce time to close?
Finance teams are using AI to automate the most manual parts of the close process, including formula generation, data reconciliation, and workflow scripting. According to SpendHound’s 2026 CFO AI Report, which is based on interviews with 30+ CFOs, teams are using tools like Microsoft Copilot, ChatGPT (OpenAI), and Gemini to generate spreadsheet logic, debug errors, and iterate on models more quickly. This reduces the time spent maintaining spreadsheets and allows teams to move faster from data assembly to analysis.
What AI tools are finance teams using for close automation?
Finance teams are primarily using general-purpose AI tools that integrate into their existing workflows. The most common tools cited in SpendHound’s 2026 CFO AI Report include Microsoft Copilot for Excel-based workflows, ChatGPT (OpenAI) for scripting and automation, and Gemini for use within Google Sheets and Workspace. These tools do not replace ERP systems but instead help automate the manual work required to connect and manipulate data across systems during the close process.
How does AI help with financial close workflows?
AI helps by reducing the time required to build, maintain, and debug spreadsheet logic. Instead of manually writing formulas or scripts, finance teams can describe what they need in plain language and have AI generate a working version. This makes it easier to handle complex revenue recognition scenarios, iterate on models, and resolve errors. As a result, close workflows become more repeatable and less dependent on manual intervention, leading to shorter and more predictable close cycles.
What results do finance teams see from using AI in the close process?
According to SpendHound’s 2026 CFO AI Report, finance teams using AI in close workflows report faster close cycles, reduced time spent on spreadsheet maintenance, and fewer manual errors. Teams are also able to handle more complex reporting requirements without adding headcount. One CFO noted that tasks that previously took over an hour could be completed in minutes using AI-assisted scripting and formula generation.
What needs to be in place before using AI to accelerate the close?
Three conditions typically need to be met before AI can effectively improve close workflows. First, the underlying data needs to be reasonably structured and accessible, even if it lives across multiple systems. Second, finance teams need to identify the most manual and repetitive parts of the close process, such as formula building and data reconciliation. Third, workflows should be stable enough that automation can be applied consistently across periods. Without these foundations, AI application to the close process may not speed up the process or improve accuracy.
Where does SpendHound’s spend data come from?
AI tool pricing varies depending on usage, company size, and contract structure. Based on SpendHound’s SaaS pricing benchmarks—derived from de-identified spend data across 1,000+ companies—ChatGPT (OpenAI)costs approximately $42,000 annually for SMBs (50–999 employees) and $167,000 for enterprise companies (1,000+ employees). Microsoft Copilot and Gemini are typically bundled with Microsoft 365 and Google Workspace, respectively. Note: These numbers were accurate as of the time of the publication of the full report. For the most up-to-date real-time prices for ChatGPT (OpenAI) and other AI and SaaS software tools, check out our Marketplace.
Can AI replace the financial close process?
No — AI does not replace the close process or financial judgment. Instead, it supports finance teams by automating repetitive, rules-based tasks and reducing manual effort. The CFOs interviewed in SpendHound’s 2026 CFO AI Report consistently emphasized that human review remains essential, especially for validating outputs and ensuring accuracy. AI is most effective when used to accelerate workflows, not replace decision-making.
Why do finance teams still wait until the end of the month to analyze data?
Even when finance leaders want more real-time visibility, the effort required to pull and reconcile data mid-cycle is often too high. Data typically sits across multiple systems, and assembling it manually can be time-consuming. As a result, many teams defer analysis until the close, when data has already been consolidated. AI helps reduce this friction by making it easier to generate formulas, connect data, and analyze results without waiting for the close. As AI tooling matures, more finance teams are beginning to pull mid-cycle data on demand rather than deferring everything to month-end.
Report findings are based on 1:1 interviews with 30+ CFOs from December 2025 through February 2026. CFOs included in the study represent both SMB and enterprise companies ranging in size from 11 to 5000+ employees. They represent experience across a broad range of industries that include technology, financial services, software development, advertising, healthcare, education, hospitality, advertising, and travel. Some CFOs have been anonymized for the purposes of this report.
For information about our spend insights and pricing benchmarks, check out our full methodology.
Book a demo below and we'll get you set up with our team.
