The Truth About How AI Can Help Early-Stage Companies

Introduction

Since OpenAI released ChatGPT, the business world hasn’t stopped talking about AI. It has become the ultimate buzzword, often used to grab attention. But what does AI really mean, particularly when it comes to corporate finance functions?

We’ve all seen blog posts and videos proclaiming that AI will revolutionize finance or even eliminate some roles entirely. And sure, maybe some of those predictions could come true years down the road. But what about the here and now—or the next five years?

What role can AI realistically play for most finance teams, especially those working in early-stage companies?

Manual Processes and Errors

It’s no secret that even today, many FP&A teams are still using processes and spreadsheet techniques that date back to the 1990s. Why, you might ask? Because finance teams know that, even with a lengthy and clunky process, they can still achieve the needed end result—and that’s all that matters.

We’re still far from the day when a finance executive will feel comfortable handing over a fully completed forecast to an algorithm, especially one that can still produce errors and lacks the ability to think critically.

What AI can do, however, is assist with the aggregation and transformation of data before it’s loaded into Excel or a business intelligence tool for summarization and analysis.

You can ask AI to review a dataset and validate specific items, such as ensuring columns don’t have missing data, identifying duplicates, or summing up the total in a particular values column. These are all useful tasks that can help streamline your process while still allowing you to maintain oversight of how everything is being run.

Forecasting Capabilities

No one is going to trust AI to run a complete forecast on its own. Many businesses use a bottom-up approach with numerous moving parts, making it challenging for AI to fully handle the complexities.

However, AI can be valuable in scenario analysis for tweaking assumptions. For example, imagine you have a dataset of current customer contracts with renewals scheduled for the upcoming year.

In one column, you have the renewal amounts, and in another, the renewal percentage you’re assuming for each customer by region (e.g., USA or Europe). Let’s say we’re assuming an 80% renewal rate across all customers in 2025, which would result in an ARR of $30M at the end of the year.

We could then ask AI: “What would the ARR be in 2025 if we assumed an 85% renewal rate for North American customers and a 75% renewal rate for European customers?” AI could quickly apply this logic to the dataset and provide the result—much faster than manually copying and pasting formulas and filtering results.

You can use AI to evaluate these assumption adjustments, but in the end, it’s up to you to update the model correctly and ensure the assumptions hold true.

AI can help guide us to the right answer more quickly, but it’s ultimately our responsibility to perform the final calculations and validate the recommendations.

Focus on Data Quality

AI can assist in evaluating assumption adjustments, but it’s ultimately your responsibility to update the model accurately and ensure the assumptions are valid.

While AI can guide us to the correct answer more efficiently, the final calculations and validation of recommendations rest in our hands.

Conclusion

Do not get lost in the AI buzz, yes there are many day to day uses for it that can help us be a bit more productive but we still have a long way to go before it will completely automate the task that FP&A team will be doing on a day to day basis...