Start-ups, algorithms and LLMs for lawyers and accountants. Where do we stand?
A growing number of artificial intelligence software is now targeting the world of accountants and lawyers. Here is a brief guide.
5' min read
5' min read
Automation of repetitive processes - accounting, tax, legal. Supporting the analysis of data, documents. With a double benefit: saving time and improving the quality of one's work. A growing number of artificial intelligence software is now targeting the world of accountants and lawyers. A market that is maturing, so much so that the big names of American software houses are being joined by many Italian companies. Specialised start-ups or historic names in Italian software.
Commercialists
For example, 'we use DK Mind from Datev Koinos and Team Systems,' says Fabio Grieco, an accountant at Srg Associati. The former is a 2002 company resulting from the collaboration between the Datev cooperative of accountants and the German software house Koinos. TeamSystems Spa is a well-known Italian manufacturer of management software since 1982. 'DK Mind uses artificial intelligence algorithms to translate documents, even paper documents, into accounting elements automatically,' he adds. So there is an automation of accounting, invoices. "Thanks to machine learning it learns by itself how the different invoices loaded should be accounted for; it avoids repetitive manual processes and saves time."
Another advantage, "it allows the use of unbundled accounting data to compile databases on which to make historical and predictive analyses of business activity. For example, it helps to forecast cash needs, to predict - based on payment history received and expected growth - the financial needs of a business client of that accountant. The professional is therefore more effective in supporting clients.
"TeamSystem's AI does similar things, but we mainly use it to make work on customers more efficient," says Grieco. "For example, it estimates the working time of each employee on each customer and can thus help us understand if we are devoting an incongruous number of hours to a certain customer, to the point of working at a loss, because we have assigned it to an employee who is not suited to that type of task." "Then simply assigning the customer to another employee can greatly improve efficiency."
Another possible frontier is generative AI, 'we are experimenting with some clients with the Hector software (by the Canadian company of the same name), to automate the writing of self-defence petitions, appeals; but it does not work very well yet. It manages to retrieve all normative references but the Italian tax situation is so particular that it does not allow this level of automatic generalisation,' Grieco explains.


