In the popular children’s book Charlie and the Chocolate Factory, the main character’s father has a job screwing the caps onto toothpaste tubes. Unfortunately, the factory he works at replaces him with a machine. In short, automation technology has displaced him.
Automation is often traditionally associated with physical processes. However, automation in finance contexts can complete financial and back-office tasks (examples being data extraction and analysis, validation, underwriting and more). These finance processes are more contextual and nuanced than screwing on toothpaste caps. A successful automation solution in finance relies on more than just mechanical repetition – it requires the automation tool to be able to ‘read’ the information like a human.
Automation technologies usually incorporate deep learning algorithms that can comprehend how financial information interacts in a given context. For example, an intelligent automation tool could not only identify the obvious indicators of credit risk but also delve into the subtle nuances of a borrower's financial situation, much like a seasoned credit analyst. Plus, it’s consistently faster and cheaper than human labour.
With the increasing sophistication of automation technology in mind, it’s understandable to wonder about the limits of such technologies and whether it means job displacement is a possibility. It turns out this has always been a concern throughout the history of automation in finance.
Financial automation technology kicked off with the invention of the Pascaline, invented by Blaise Pascal in 1642. The Pascaline was a mechanical calculator capable of simple addition and subtraction. While it required manual operation, it mechanised the calculation process, paving the way for the development of financial automation tools.
Another notable contribution occurred in 1801 when French inventor Joseph Marie Jacquard created a loom that used punched cards to automate weaving. Jacquard's invention inspired the development of early computers, which also employed punched cards for data input.
Fast forward to the 1950s, when bookkeeping changed drastically. Instead of punched cards and batch payments, bookkeepers turned to rudimentary computerised technology. Electronic Recording Machine Accounting (ERMA) laid the foundation for automated cheque handling and credit card processing. Around the same time, early Automated Teller Machine (ATM) models were being developed. In 1967, three companies launched versions of ATMs that paved the way for more sophisticated models.
Later in the 20th century, spreadsheets entered the scene. In 1969, Rene Pardo and Remy Landau invented LANPAR, a spreadsheet exclusively available for mainframe computers. In the 1980s, spreadsheets became more accessible due to the rise of personal computers. VisiCalc and Lotus 1-2-3 are early examples of spreadsheet programs.
Automated finance technology in the 20th century offered two significant advantages compared to its predecessors. Firstly, it allowed finance professionals to ditch repetitive tasks like recording financial transactions. It also empowered customers to manage their finances more autonomously. By 1967, three ATM models had been released globally, meaning consumers could access cash without visiting a bank.
One of the most significant milestones in financial automation technology occurred in November 2022 – the release of ChatGPT. ChatGPT could seemingly do it all – calculate key financial ratios and produce forecasts from inputted data. Note: Upon release, it could not extract data from documents, as it lacked the functionality to add attachments.
The term ‘AI’ has assumed a different meaning in the last two years. Its starring role in financial services has meant that – without exaggeration – financial technology will never be the same again. Let’s explore this further.
Instead of focusing on what automation can do, conversations have shifted towards what it can’t do. Concerns over obsoletion have spiralled, with tests being conducted for soon-to-be automated entry-level positions in firms like Goldman Sachs and Morgan Stanley.
Automation excels at back-office administrative tasks that entry-level associates may be expected to do (or more experienced professionals forced to do). Currently, finance firms can reliably automate several key processes, including:
A prime example includes automatically extracting the line items from invoices.
One use would involve calculating financial ratios (like operating expenses) from balance sheets.
Intelligent algorithms can structure information by reading ‘documents’ to make or recommend loan decisions. Some AI technologies can then generate loan underwriting documents, too (an example of document automation).
When implemented conscientiously (i.e. using a stable connection, well-calibrated automation technology, etc.), automation technology offers several benefits for organisations of all sizes. For example:
Of course, these statistics are all estimates. Gathering precise data on the effects of automated finance processes can be difficult. Data privacy concerns and the intangible nature of certain benefits, like improved decision-making, make it challenging to quantify the exact success of AI.
However, we discovered that our AI-powered automation:
You can discover more statistics about the results of Evolution AI’s automation in our Case Studies.
Automation’s undeniable (and tangible) benefits mean its present state is akin to a rat race with developers racing towards releasing automation technologies. Though automation can offer short-term benefits (like reducing tedious, repetitive labour), it is also a long-term investment for finance firms.
Of course, automation processes will differ from sector to sector. Let’s examine how organisations in one sector are deploying automation.
In 2024, Gartner noted that 73% of accountants were experiencing increased workloads as a result of new regulations. Accountants play a crucial role in safeguarding financial health through meticulous procedures and analysis. The result is that many accountants have to then complete tedious tasks alongside sophisticated analysis and exercising their professional judgment.
Automation technology is evolving to alleviate and ideally remove the admin burden. Many manual processes in accounting can easily be condensed into automated processes. For example:
Sophisticated AI data extraction algorithms power many of these technologies that extract data from sources (like financial statements and invoices). So it can be stored or used to trigger the next action (e.g. responding to a customer, creating an invoice and more).
As a result, the role of the traditional accountant is changing. For example, accountants may have more time for client communication. Clients appreciate speedy and proactive communication, which may not be possible if all of an accountant’s available work time is spent managing funds. Some accountants may also pivot towards other strategic technological initiatives, such as AI financial modelling.
Therefore, automation in accounting is flexible and diverse. There is no need for accountants to waste time filing documents, copying data into spreadsheets, matching payments and so on.
So far in this article, we’ve made several references to ‘AI-powered automation’. Automation has taken huge strides alongside advances in AI (notably advances in LLMs, showcased by ChatGPT’s debut).
Intelligent Automation (IA), also known as Intelligent Process Automation (IPA), refers to the use of AI and machine learning algorithms to enhance the performance of robotic systems. Rather than repetitively following rule-based algorithms, intelligent automation adds flexibility to automatic processes. AI can review the information contextually, linking it to external databases where necessary, to make informed decisions. When manually corrected, AI will not repeat the same mistake (compared to old-school automation, which required extensive retraining to avoid repeating errors).
Therefore, intelligent, or AI-powered automation works directly alongside finance professionals. AI can adapt with minimal effort as the market conditions, a company’s trajectory and clients evolve.
While the future trajectory of automation in finance is not agreed upon (either by financial institutions or the fintechs working alongside them), there are a few potential paths that automation could take. Let’s explore a few scenarios.
Affordable, pre-configured automation solutions mean that global companies of all sizes can automate simple processes. From the consumer’s perspective, successful automation means faster services. No more painful mortgage approvals, waiting for funds to clear, time-consuming customer service, and so on.
Wider access to automation tools could also improve financial automation software’s performance, leading to…
The increased uptake of automation tools renders the automation landscape more competitive than ever. Vendors are incentivised to add extra functionalities, as well as improve their software’s performance. What does this look like in practice?
Increases in computational power mean that automation technology will become lightning-fast.
Generally, financial automation tools deliver a reliable performance. New advances in AI must deliver consistently reliable performance. Vendors should ensure bugs are fixed, and data loss is preventable, alongside other measures of reliability.
As professionals become more familiar with AI and its capabilities, they will be able to integrate it seamlessly into their workflows. This synergy between human expertise and developing AI technologies will lead to the most efficient processes.
As more companies invest in financial automation, the most probable outcome is that the performance of these tools will increase exponentially.
At the same time, improved output may spark fears of obsoletion among finance professionals, who worry this could potentially lead to end-to-end automation for all finance processes.
While promoting partial automation, McKinsey, perhaps naively, suggests that employees made obsolete through automation will be displaced to other roles (tell this to entry-level staff). However, they make a helpful point – that productivity gains ultimately occur by interacting with humans.
Automated finance processes seem to tip the dynamic towards senior professionals working alongside intelligent automation. However, it’s important to ensure that less experienced but more tech-savvy professionals are not phased out (particularly when financial sectors like insurance are desperately understaffed).
If automation technologies wholly replace juniors, it can have severe ramifications on the future of financial services. The best way to make the most of AI productivity gains is to get everyone involved.
As this XKCD comic explains, developing an automation tool can consume the time freed by automating the task. Companies like JP Morgan can use their considerable resources to develop their own automation tools quickly. But, for many companies, this isn’t an option.
The way to escape the development/time conundrum is to use an external tool. Transferring the responsibility of creating, testing and continually improving the tool to a vendor can unlock the benefits of automation. But you’ve got to find the right vendor.
(We’ve written extensively about finding the right vendor – check out our cheatsheet here).
Of course, the best way to navigate the increasingly complex automation landscape in the finance world is to try the tools yourself. Sign up to try them out, attend demos and then ask the right questions. Researching financial automation tools can be a challenging process, but knowing you can experience a high ROI is a powerful incentive.
Evolution AI has worked alongside global leaders like Dun & Bradstreet, Hitachi Capital and NatWest to cultivate high-performing document automation products.
Evolution AI is proudly powered by AI expertise. Book a demo to find out more.
