Odoo launches AI accounting assistant for invoice and data analysis

Hello everyone. My name is Yiannis. I'm an accounting expert at Odoo and today I'm going to present you the new accounting assistant powered by AI. Please feel free to scan this QR code if you want to ask any question. I will review them at the end of the talk. So, we all know that accounting is one of the most important aspects when running a business. But how quickly can you turn the information into decision? Whether you are an accountant, a business owner, or a business leader, you need clear answer. You need to know what what need attention, what has changed and what should happen next. These expectation are driving a revolution and today we will explore what that means for your everyday work. So traditional accounting uh has always follow a familiar rhythm. You need to record the transactions, search manually manually for the information, and prepare the reports for compliance. But this task remains still essential but finding and interpreting those results can take time. This is where why the accounting is changing now because AI allow us to ask directly for the information without to have have to know which menu or which filter to use, analyze continuously the data, and act while the data is fresh. So it can really impact your decisions. So today we will review four level of assistance. The first one will be how we can navigate through the data thanks to the AI. Then we will go on to how the assistant can help us into the uncoding of of everyday work. We will also see how it can help us to control our data and correct it. And finally, after everything is clean, how it can make comprehensive analysis and give us advice. So let's make it concrete. Let's first start with a simple use case. Let's say that I have a cash meeting this week and I need to see all my overdue invoice that are more than 30 days. I don't do not want the list of all my unpaid invoice. I just want the one that I too late for me. So I already prepared some prompt so we can focus on the accounting results. So the first one will be to show me all the invoice that are due for more than 30 days. I want also the AI to be able to retrieve more information in the sense that I want the customer, the invoice number, the due date, but also the number of days that is overdue. So, as you can see on the screen right there, is working behind and I can see all the step that is taking. I can see and understand the logic behind what he's doing and then we have the results. So, if I extend this uh this window, we can see that he has retrieved the data in seconds. So, what is really interesting here is that is not only retrieve the store data that you can find here. He's also calculating new data out of the data that is existing in the database. Okay, that's great. Let's go further. Let's do another use case. So, this time I would like to group all my vendor bills by expense. So, this is what I'm going to do right there. I'm going to just ask him to group all my vendor bills of 2026 by category. So, I open the AI and I ask him this. So, basically here is not only looking at the data, is also looking at each vendor lines and analyzing for each vendor lines the expense account that is behind that. When he find it, then he can gather it in a table and show me uh the table that is ready to take action. So, this is really interesting in the sense that not only for each category is able to do that, but also it allow us to not search for the appropriate menu and locate eventually the appropriate filter. Which this is really what matters in AI is like with an everyday prompt, we can ask in everyday language uh how the AI can help us. So, here you can see that now I have my uh purchase, my expenses grouped by the uh expense, and I can really drill down easily into what drive my costs, okay? So, this is a calculated by the AI. This is not something that I have. So, it really retrieve the data, gather it, structure it, and give me more advice. As you can see here, I also have a list of the biggest vendor bills, so I can easily take action uh onto uh these uh partners. Okay. So, this is only retrieving and navigating through the data. Let's see now how the AI can assist us into processing this information. A lot of um accounting teams spend times retrieving the data from a customer, from documents, or from another system. Here, we will see, for example, the use case where we need to import the trial balance of a customer and thanks to a PDF. So, what I'm going to ask to the AI is to import me the opening balance of this uh customer, retrieve me all the accounts, and create me the opening entry. Okay, so I open the AI, I ask the prompt, and then I can join my document that I want to import. Great. So, as you can see here, is Odoo will not just read the data. He will check if the data is correct. He will check in each of the lines of my import balance if everything is balanced, if the account are already present in the chart of account, if he need to create new one, or if uh it spot any mistakes. Okay, in the meantime we can ask another uh another uh request while it's charging. So, what I would like to do as well to help my encoding is to check in my draft invoice if I use the right account. So, basically I receive a invoice through PayPal uh from one of my uh my supplier and I'm not really sure which account to use. I know that it should be an asset, but I'm not sure which account to use and how to create the asset. So, that's what I'm I'm asking here. I'm actually telling him to look at my purchase invoice, my draft one that I received from PayPal, and to look if I apply the right account or not. So, first it will retrieve all the invoice, retrieve all the account, retrieve all the product, and check if everything was correct. So, coming back to my first use case, you can see here that the account the AI is not only uh looking for data. It's only also asking me to validate the different step. So, it's not taking decision unless I'll tell him. I trust him for this time and I will always approve what he's asking me uh today. So, what is also interesting in the AI here is that you can really follow every step. You can see which record is looking at. You can see the logic behind and you can understand also if it made mistake or not. So, you can spot really what he's doing. This help us actually to gain a lot of time. You know, um an an import balance can have sometimes dozens, sometimes hundreds of lines. Doing it manually uh can be a big source of mistake and also of course is time consuming. So, let's come back to my opening balance and from here you can see that I have all my discussions. So I can actually ask different requests at the same time to different agents. So this one is particular for the accounting, but I can of course add skills, I can add context so the answer and the result will be really specific to my needs. So as you can see here is still looking, sorry. Now is looking at the draft invoice and looking also at my assets settings. So in my accounts I decided to put some asset settings in and some asset models and this is what is doing now. Looking first at my rules that I decided in my database and then matching it with the actual data that is coming from there. So there is really a shift between you the work of the accountant. It's not only helping you to move faster, it's also advising you. So you can see here that actually identify that one of the purchase had an IT equipment. I asked him to if if there is an IT equipment, please please spot it and change the account. So for example here is is spotted that actually I had an expense account but I needed to have assets account. Okay? So what I'm going to say, I'm going to say okay correct the lines with the right account. So now I'm giving him the permission to correct it. In the meantime coming back to my trial balance, you can see that it retrieved all the data directly. So you can see that all the accounts were there. It didn't find any missing accounts and now I access my draft entry right there. So, I can really see now the impact of the AI within my database. So, this works for importing trial balance, but this can also work for other documents like uh, loan reimbursement tables or even asset depreciation schedules. So, you can see that you can really gain time and reduce the mistake uh, by using uh, your assistant. So, I will post the entry because everything is correct. You can see that it's balanced. And at the same times, I can review also the other prompt is also asking me if he can change the account again. I going to say yes, and now he have already updated the records. So, I can click on that, so I can audit that, and you can see that now the accounts for the the two product that I had in my purchase uh, was changed to the uh, to the right one. So, if I click on the journal entry and I post it, uh, if I post it, you can see that now the model the asset models is linked to my vendor bill. If I post it, it will create me automatically the assets. And now I can review the depreciation schedules right there. So, in minutes, you can really import and transfer the data between system or documents. Okay. So, now that we know how to retrieve the data, how to help us assist into uh, the encoding and the everyday work, let's see how the AI can uh, help us control the data and correct. So, the next use case is about um, look it uh, looking at my VAT treatments. So, basically, I have some purchasing invoice from vendors with beverage and specifically alcoholic beverage. I'm not quite sure which tax which tax I should apply to it. Um, so I will first ask the AI for his advice and then spot any any discrepancies about that. Okay. So the goal here um is actually to looking at all my vendor lines. First, then check the tax that is applied, then check also the local regulation. So it's not only looking at the data, it's comparing with outside information. When it find the data, it can now after that