As far as coverage ratios there’s not much to say. You’ll see a certain stability, only small effects but absolutely stable. In other words, the bank’s balance sheet continues to be very well covered and protected with very adequate provisions. As for non-interest expenses, this quarter is typically subject to stronger effects, such as the collective wage agreement. As a result, personal expenses accelerated from the second to the third quarter, while other expenses are very much in line. Thus, non-interest expenses grew 8.4% In Brazil, and with the effect of Latin America, they grew 6.9%. And what are the key messages? The efficiency ratios are quite good, clearly much better than what we’ve seen in the market as a whole both in Brazil and on a consolidated basis.
And these are international benchmarks. We’ve managed to deliver a very appropriate efficiency ratio with two main messages. The first is about the bank’s core cost or run the bank, which is in line. In the nine-month period compared to the same period in 2022, we grew only 1.1%. On the other hand, what has actually been expanding this figure is not just cost itself, but the investment that we continue to make [Indiscernible] our goal is not to manage costs for the quarter. Basically, what we have to do is to make our operation more productive, more efficient, thinking about how we invest in our operation by investing much more in technology data and business expansion. So we’re always looking at the franchise over the long run, always with a longer time horizon.
So that’s the reason for all these investments, which is still being absorbed by the P&L resulting in the level of profitability I’ve just mentioned. I believe these are the key messages regarding costs, the bank’s efficiency program continues to make a very positive contribution. And in terms of transactional volumes, if the unit cost is the same as lower or rising less than inflation, if we actually increase volumes to do more business, this is a benign cost, so we’ve still been able to finance all this benign cost expansion with all the efficiency program at the bank. One of the most talked about topics lately is data. We’ve talked a lot about machine learning models, generative artificial intelligence, among others. This is a topic that comes up all the time.
So what we wanted to do here was to provide a summary of our various initiatives. This isn’t just a topic for a specific department, it’s a topic for the whole bank. And we have some data that shows and reinforces how strong our investment and belief in this data agenda has been. Starting with our data structure, which has 100% of all the bank’s data in the cloud, in a very modern data mesh architecture, which makes the data much more democratized within the institution not being used by just one department as all departments start consuming that data. And not just consume, but adding their own data much more efficiently to the bank as a whole. And so we brought you some information that I think is relevant. We have more than 350 data scientists in the organization, more than 200 initiatives using generative artificial intelligence, more than 50 machine learning engineers, more than 150 professionals working with a generative artificial intelligence, and more than 570 models currently being used within the organization.
One of the cases that I think is relevant in terms of outcomes rather than output is service. For example, we’ve increased by 45 percentage points, the volume of client service that is automatically retained through our models using artificial intelligence. 72% of all calls made already handled by artificial intelligence with much greater efficiency, accuracy and speed, and with improved NPS. And this is in line with all the investment in technology and efficiency that I’ve just mentioned. This shows a much more scalable and efficient bank in the long run. And we have a series of other initiatives with greater security for our clients. Since we are able to interact and identify the voice of a fraudster thus allowing us to protect our clients.
With regard to productivity and the corporate client experience, we already have a lot of information for every documentation analysis so that it can be done as accurately as possible. We currently have a 97% accuracy. We’ve also been using chatbots to interact with our clients, providers. We’ve used our artificial intelligence models in different businesses, and we have no doubt this will be an agenda that has come to stay and will grow exponentially over the coming years. We want to be at the forefront, we have no choice but to migrate our systems to the cloud and upgrade them. As for the artificial intelligence agenda, we have everything it takes to lead this process, we want to be at the forefront of this agenda. I’d also like to comment on a few topics about culture and people which are very important to us.
