Artificial Intelligence (AI) and Machine Learning are increasingly revolutionizing the world of accounting. In this article, Arthur Wallerco-founder of Pennylane, explores how these technologies are transforming the accounting profession and specifies Pennylane’s role as a key player in this evolution.
The Various Use Cases of AI in Accounting
It is important to note that Artificial Intelligence (AI) and Machine Learning are not new concepts. For years, companies have used these technologies daily to increase their productivity. A well-known example is the use of optical character recognition (OCR) technologies to process documents, a practice that has become more widespread in the last decade. However, what is revolutionary is the emergence of Large Language Models (LLM), with ChatGPT at the forefront.
At the moment, Pennylane is already putting AI models into production for several use cases, including:
OCR on purchase and sales invoices – Initially, we outsourced OCR to experts such as Mindee AND Klippa. However, due to our growth, we developed our own OCR internally, providing greater accuracy and speed thanks to the huge volume of invoices processed and feedback from our users. Our OCR now processes 200,000 invoices per day with 91% accuracy, averaging 5.34 seconds per invoice from receipt to processing.
Accounting record – For third-party invoices, Pennylane uses machine learning models based on the file’s FEC to recommend a ledger account in exchange for the 401 account. For bank transactions, machine learning models are applied to recommend a ledger account in exchange for the 401 account …count 512.
Invoice <> transaction reconciliation suggestion – Pennylane offers reconciliation suggestions based on characteristics such as amount, date, text, invoice number, and more.
Suggested analytical categorization – Pennylane suggests analytical categories based on the text of a transaction or the elements of an invoice.
Next step: Smart Review
The next step we are considering is Smart Review, which involves applying machine learning to review alerts and save time.
However, I would like to emphasize that I am not convinced that LLMs will bring about a major revolution in accounting production. Accounting is based on very structured data and I think that the gap between existing automatic rules and LLM will not be as significant as in other domains where data is less structured, such as the legal domain.
In my opinion, the real revolution in AI and LLMs is in the area of consulting for business leaders. With the advent of electronic invoicing, the central role in creating a company’s value naturally shifts from the area of production and declarations to consultancy. Companies will have to prepare their employees for these new missions. AI and LLMs represent an extraordinary tool in which the profession’s editors must invest massively to increase the company’s headcount and help them fulfill new missions with high added value for their clients. In concrete terms, this means that it will soon be possible to generate tailor-made quality reports in record time, with the help of AI, to retrieve and format data.
Pennylane: a key actor in this evolution
One of the main criteria for choosing a partner publisher in this new environment is that this publisher has interests perfectly aligned with yours. At Pennylane, we choose to work closely with accountants, offering an indirect distribution model that does not compete with companies, unlike other players who directly target their customers. Our business model ensures that our interests are always aligned with yours.
Furthermore, Pennylane is an independent and sovereign French company, controlled by its founders and employees, all of whom are shareholders. We systematically share with you the value generated by subscriptions purchased by your customers, to align our interests.
When it comes to technology, Pennylane is investing heavily in the data science space. We have a strong and experienced team of 25 people, including data engineers, machine learning engineers, data science product managers, and data analysts. Since the beginning, we have invested 3 million euros in building our team and our data infrastructure. Our commitment in the field of data will exceed 3 million euros by the year 2024. Our development of internal OCR technology and our standardized database for third parties attest to the solidity of our infrastructure.
Regarding data security and sovereignty, our T&Cs are clear: the company remains the owner of the data. We will never share your data with other companies or third parties unless you give us permission. We are ISO27001 certified and constantly invest in cybersecurity protection.
Ultimately, Pennylane offers a complete solution that allows you to equip your customers with a management tool while maintaining your proprietary ecosystem. This helps prevent your customers from managing things outside of your environment, which can make it difficult to access their data in real time or in the correct format. Plus, it gives you the opportunity to offer new services to your customers more easily.
In conclusion, AI and LLMs are transforming the accounting landscape, changing the value center for business consulting.
To learn more about this subject, Arthur Waller invites you to a discussion on: AI and accounting: what impact on companies? , Thursday, December 14, at 11:30 am.