How AI is revolutionising finance processes

The world of finance has long adapted to technology, from calculators to Excel spreadsheets and cloud computing. Now, artificial intelligence — especially generative AI and machine learning — is at the helm of accelerating change.  

A January 2025 report from Bloomberg Intelligence concluded that in the next three to five years, AI will be able to carry out back office, middle office and operations tasks for global banks. This will lead to over 200,000 jobs being replaced, with Tomasz Noetzel, the senior analyst who wrote the report, stating that “AI ill not eliminate [these jobs] fully, rather it will lead to workforce transformation.” 

The peer group included JPMorgan Chase & Co., Goldman Sachs Group Inc. and more, with respondents expecting an increase in earnings over the next few years. According to the report, eight in ten agree that generative AI will increase productivity and revenue generation by at least 5% in the next three to five years. 

 

But, like every major technological shift, challenges arise. From how to integrate AI into existing tools and workflows to tackling budget constraints, organisational cultural resistance, and a growing skills gap, questions surrounding the practical applications of AI are only getting bigger. It’s no surprise the World Economic Forum is already branding the finance sector as one of the most vulnerable industries to AI disruption.

The peer group included JPMorgan Chase & Co., Goldman Sachs Group Inc. and more, with respondents expecting an increase in earnings over the next few years. According to the report, eight in ten agree that generative AI will increase productivity and revenue generation by at least 5% in the next three to five years. 

But, like every major technological shift, challenges arise. From how to integrate AI into existing tools and workflows to tackling budget constraints, organisational cultural resistance, and a growing skills gap, questions surrounding the practical applications of AI are only getting bigger. It’s no surprise the World Economic Forum is already branding the finance sector as one of the most vulnerable industries to AI disruption. 

And yet, its potential is undeniable. AI has been shown to cut company costs, boost workforce productivity, and make processes far more efficient. There’s no question that the AI financial revolution is well underway.  

But how can you actually put AI to work in your finance processes? This was recently discussed at Forge Connect, an event Rydoo and iBanFirst hosted that brought finance professionals together in Amsterdam to discuss leading industry trends. Nicolas Boucher, one of the guest speakers, is an AI trainer for finance professionals, having worked in finance for over a decade before this.

Boucher understands the value of technology and finance better than many, having trained more than 5,000 professionals from companies such as KPMG and Mercedes Benz on how to use AI for finance. At Forge Connect, Nicolas shared his proven formula for applying AI to finance processes — and the results are nothing short of eye-opening.  

When generative AI first emerged, many underestimated the tool, thinking it was mainly used to write content and retrieve static information. Nobody talked about it for business, and even less for finance. Fast-forward to today and AI has become an indispensable assistant, helping finance teams complete white-collar tasks like writing, problem-solving, and creating frameworks in minutes instead of hours. 

For Nicolas, the leaps in AI are extraordinary. “When comparing departments, you can now generate intuitive, insightful graphs in moments instead of hours, and the accuracy is unmatched”, he explained. “We all remember the time when you wanted to see internal controls, and you knew that they didn’t exist. So you had to stay up late or work with the team to do that. I remember having to do that myself. And now GPT drafts these procedures for you.” 

Today, around 58% of finance professionals are using AI for their daily tasks, up from 37% a year ago. The most common use cases include using AI tools to analyse a large volume of data, generate reports, automate processes like business expenses, and improve customer and service delivery. What’s more, over 80% of finance professionals report that AI is already boosting revenue and lowering annual costs, propelling its usage across the industry.

Finance thought leaders are suggesting that the next level of transformation with AI is underway, with AI slowly starting to integrate into existing systems like ERPs and data platforms. While AI is becoming increasingly smarter at accessing and analysing data, most companies are only now beginning to explore this potential, with a big learning curve ahead of them. Indeed, top industries needing a skills upgrade as a result of AI include the financial services sector. 

According to Nicolas, the focus today is not on whether to use ChatGPT — everyone does — but rather on learning how to use Python tools, one of the most popular programming languages, where users can build software, develop web services, perform data analytics and even train machine learning models for free. Due to its nature, Python tools can overcome many of Excel’s limitations, especially when it comes to data visualizations, data processing, and superior forecasting. In short, it enables finance teams to analyse large volumes of data and automate tedious, time-consuming tasks.  

It was precisely at a moment where he needed to analyse multiple data files for an audit report that Nicolas turned to this programming language. “We needed to upload 80 files with 100 lines each for the audit, and the accounting software couldn’t accept the files as they were”, Boucher recalls. “It needed debit and credit columns, some lines had to be removed. It would take a week, and meant high IT costs to do it.” 

Nicolas’s solution? He turned to Python and managed to merge dozens of files together in under a minute. After a few iterations with GPT to refine the code, Nicolas had a finished file, which he sent to his accountant to check if everything was clear. Now, he does this same process every quarter in just a few minutes.  

It’s increasingly clear that the evolution of AI in finance is both rapid and transformative, with tools such as Python enabling finance teams to leverage AI within their own environments. This means saving significant amounts of time and automating highly complex tasks in ways that seemed impossible just a year ago.

Finance teams often spend a lot of time manually reviewing expense claims to ensure they meet company policies. Faulty receipts only add to their workloads since every transaction must be checked. And with nearly half of employee dishonesty cases involving such practices, the process of combing through expenses is not only time-intensive but prone to mistakes. The bottom line is: it’s impractical.  

There was a time when Nicolas had to take over a team of accountants that had been recently downsized. And while the team got shorter, the pile of employee receipts continued to grow. 

“People started complaining that their expenses were not validated”, Nicolas recalls. “Nobody wanted to do that task. Today, AI is doing the tasks we don’t want to do.”

For businesses, especially those that deal with multiple expense claims, a new era has cropped up. One in which teams can be more efficient, precise, and productive. From automating routine tasks to providing insightful data analysis in real-time while minimising manual effort, AI is gradually becoming an ally for finance teams, who now have the time to focus on value-adding tasks. “If you want to create value and have a better career, you need to be better at human tasks and manage AI to do the robotic tasks”, Nicolas concluded. 

Rydoo’s Smart Audit model is a great example of this. The tool uses AI to automatically review and monitor all expense claims to catch even the smallest non-compliant details that could easily slip through the cracks. This means fewer fraudulent claims approved and a lighter load for both approvers and finance teams, giving them more time to focus on strategic and impactful tasks. 

CFOs no longer need to speak to several departments to collect their insights and predictions and then compile them into static forecasting reports. Today, AI and advanced data analytics have replaced that competency, allowing finance teams to focus more on financial storytelling and strategic insight rather than manual input.  

Indeed, machine learning is transforming how finance teams operate, moving them beyond analysing past data to predicting what’s ahead. By uncovering patterns in historical information, organisations can easily forecast trends, plan budgets, and allocate resources more effectively.  

For forecasting, Nicolas Boucher recommends starting with existing models such as Prophet, from Meta. When it comes to other tasks and to save time, Boucher suggests that finance professionals create their own models with AI, which can then be adapted to their businesses and finetune by adding more instructions and documents. 

“Start by creating your own chatbots that will answer all of the questions from other departments, even if they are client-facing”, he explains. “You can then capture all the questions and insights in one go.” 

Major corporations like Shell are already using machine learning to improve their forecasting. AI helps them predict shifts in energy markets, from price changes to break in demand. By understanding global trends in energy consumption and pricing, they can make more accurate revenue predictions and ultimately solve business problems on the go. 

Another added benefit of predictive analytics is its ability to spot anomalies in data and potential fraudulent activities. Whether it’s unusual spending patterns, unexpected drops in revenue, or potential risks, AI-powered models are adept at bringing these anomalies to the surface in real time. 

For finance professionals, this means quicker response times to potential issues, arming the leadership team with market insights they need to make proactive business decisions. Interestingly, a 2023 study found that organisations using AI-based fraud detection techniques reported 50% lower fraud losses and 60% faster fraud-detection compared to those not using AI. As AI algorithms advance, the accuracy and efficiency of data analysis will only get more precise.  

Building the right skills is key for finance teams to truly take advantage of AI and stay ahead in a competitive market. As teams learn to adopt and work with AI tools, these changes will naturally reshape how the organisation operates. The process starts by understanding where the finance team stands, then focus on developing their skills and, finally, turning AI into a natural part of their day-to-day work. 

According to OpenAI’s CFO, Sarah Friar, acquiring new skills to work with AI requires initiative and drive. Rather than turning to technical partners every time help is needed, finance teams should evolve their prompt engineering skills, learn how to make AI tools work into their processes, and become more active players in the business.  

“The technology has expanded the skill set of the team”, Friar explained during an interview for McKinsey’s At The Edge podcast. “In finance, it’s very useful to have someone who can write code or help with SQL [structured query language] queries, but that is not a common skill set in finance. Instead of asking for help from our technical organization, we can now just ask ChatGPT to assist in writing that SQL query. This has really advanced our team from number crunching to being a better business partner.” 

As part of this transformation, finance teams must embrace a new operational model fueled by digital skills. These include learning how to program bot algorithms, mastering analytics tools, and translating business data into actionable insights. This requires upskilling across the whole finance function.

But how can organisations effectively roll out a reskilling or upskilling strategy? First, leaders must assess the team’s existing skillset and observe how everyday tasks are completed. This will help leaders see how the team works and what skills gaps must be addressed. With this information, a training plan can be crafted in line with the team’s areas of interest. When training aligns with what employees do and where they want to go in their careers, they’re more motivated to learn and grow. This not only helps finance teams stay ahead of the curve in a changing industry but also strengthens their ability to lead and drive meaningful impact. 

The final step is to implement a structured, organisation-wide learning plan. Ultimately, a well-thought-out learning journey allows leaders to create a roadmap of skills needed to grow the business.  

While AI and finance share a promising future, it’s not without challenges. About three-quarters of decision-makers admit their organisations face significant challenges in adopting AI and machine learning. One major concern, shared by 77%, is that their data isn’t reliable or up-to-date enough to make these technologies effective. Another key issue is a skills gap, with 72% stating their teams don’t have the expertise needed to implement AI and machine learning successfully. 

This is particularly pronounced when applying GenAI to financial reporting. Concerns over transparency, privacy, accuracy, cybersecurity, and regularity clarity, all act as barriers to implementation. While companies are picking up the pace on this, the issue of confidentiality is still a major challenge. 

“What is the biggest problem when using AI with data?”, Nicolas Boucher asked during his presentation in Amsterdam. “Confidentiality. For example, when I upload data into ChatGPT, if I have a contract with Copilot, Google, etc, That’s fine if my company accepts the data security standards of the big LLMs providers, which are often similar to most cloud companies. However, many of you probably don’t have contracts with these companies to use their [Large Language Models] LLMs, so you shouldn’t use them with your real data.” 

As financial services organisations move from proof of concept to business impact, there’s growing support for continued investment in AI projects. Target areas include identifying additional AI use cases, working with stakeholders to accelerate AI adoption, optimising AI workflows and production cycles, upskilling the workforce, and investing in more sophisticated computing infrastructure. 

Certain areas of finance are already making great strides forward. For example, AI-enabled financial fraud detection is predicted to grow by over $10 billion globally in 2027. As fraudsters become more cunning in their attacks, AI-enabled detection methods will be used to fight crime.  

For finance experts like Nicolas Boucher, the future of AI in finance is crystal clear. “We should not teach our kids how to program the computer. It needs to learn how we talk and program with our own words to get the output we want. That’s where we’re going. In finance, the right people who know how to ask the right question will get further than the others.”