Lessons from Argentina on moving fast without breaking compliance - AI in CX Summit 2026 | Engageware

Lessons from Argentina on moving fast without breaking compliance – AI in CX Summit 2026

AI in CX summit | 2026

Lessons from Argentina on moving fast without breaking compliance

Jesica Cardozo [20 min]

Click for sound

●●●●●●●●●●

Introduction to Banco Colo and Digital Banking

Digital Strategy and Customer Experience Focus

Migration to Generative AI: The Journey Begins

Regulatory Challenges in AI Implementation

Customer Trust and Integration Success

Future Innovations: Money Transfers via WhatsApp

The Role of AI in Call Centers

Advice for Banks Considering AI Deployment

Benefits of AI Adoption at Banco Colo

Lessons Learned from Implementing Conversational AI

Moving fast in a regulated market is not a contradiction. It is a strategy. Jesica Cardozo from Bancor, the first bank in Argentina to deploy generative AI for customers, shares how her team scaled to 155,000 monthly sessions while staying fully compliant.

See what it covers:

The AI suite built for real business outcomes

Explore the Aivo Suite

Transcript

Download

Speakers Auto-scroll

Keyboard Shortcuts

Chapter Introduction to Banco Colo and Digital Banking

Dante Franco

Jesse, nice to have you.

Jesica Cardozo

Hello, everyone. I’m Jesica Cardozo, and I have been part of Banco Colo for the last twenty years. And today, I lead a digital banking division that is called Virtual Assistant. My journey has gone from traditional banking to the most important digital transformation in our history. We created our EIO, our main AI assistant. We call it EIO. Bancor is the name as we call it, but the official name is Bank of Province of Cordoba. It's a public institution that was founded in eighteen seventy-three. We have more than two million customers all around Argentina. And now we have two virtual assistants. But before I tell you that, you have to know that it's a very big bank. Its historical mission has always been to support local families and the productive sectors of the region. Today, it operates as one of the most important financial engines in the country. This bank, it has presence throughout the province and a convenient branch network with continuously digital channels. We have, like I say, two virtual assistants. And in EIO, we have one hundred fifty-five sessions per month.

Chapter Digital Strategy and Customer Experience Focus

Jesica Cardozo

And now we have a digital strategy that is focused on self-service, omnichannel experiences, and the progressive incorporation of artificial intelligence to improve the customers' experience. We want to be one hundred percent digital to our customers. With EIO, we last year, we have the virtual assistant in some channels like WhatsApp, web, Instagram, and Facebook. And WhatsApp was always the main channel that people chose. WhatsApp always. And we decided this year to focus on WhatsApp. And since May of twenty twenty-six, WhatsApp is the only active virtual support channel, and we are focused on self-service. We want customers to resolve their issues without interacting with a human agent.

And we have when we moved to immigration, we were migrated to a generative AI content. So now we have eighty-nine percent of sessions with active integrations. So it's our most important channel. What does the virtual assistant handle? Well, we have in one hand the transactional use cases, and on the other hand, we have general information. And with our integrations, we have people wanting to know about their credit card balance, limits, statement inquiries, purchase authorization. We have password generation of the home banking app and savings account opening. And on the other hand, we have information about loans, investments, insurance, and benefits, of course, branch hours and locations here in our province. About our transaction use cases, there, you can see two or three. The most important transaction use case are about credit card and the home banking key that is called banker password.

Chapter Migration to Generative AI: The Journey Begins

Jesica Cardozo

So people want to get personal information about their virtual assistant. Oh, migration. Okay. Well, last year, we decided to migrate to Generative AI. We started on the whole process in September. We migrated the website, Facebook, and Instagram virtual assistant first. And in December, we migrated WhatsApp. So this year, we started with full Gen AI. And this has had a lot of impact, positive impact to us. Now we can update knowledge without development, and we can edit the knowledge base and the bot response immediately. For example, we have a generative AI in our team. No code editable knowledge. So we want to be available for our customers twenty-four seven, and we want to be useful to our customers, and we are achieving it in that way.

Dante Franco

And thank you very much for that information, Jesse. We have a couple of questions that we would like to share with you. The first one is you were the first bank in Argentina to deploy AI for customers. Could you walk us through the journey of migrating your bot to Gen AI and what prompted the move? And, also, where are you today?

Jesica Cardozo

Yes. Yes. Yes. I in twenty nineteen, I started working with the virtual assistant at the bank I had recently incorporated. And this VA was conversational with low efficiency, low empathy, very poor satisfaction. So our bank aimed to incorporate new technologies that allow us to renew ourselves and go from being a traditional bank to a bank that allows us to progress and to solve the customer's queries. So in that moment, when we started with with the virtual assistant AO, we wanted to be available twenty-four seven. And we were at an inflection point because we had a VA that wasn't understanding the customer's queries. The conversational flows were so difficult to manage, and customer dissatisfaction was growing. So in twenty twenty, we decided to choose Aivo as our partner. When I say Aivo, it's a game engagement. The main reason was that we needed a platform that could scale with us. And at that moment, we needed lots of assistance that only could leave us. So Aivo’s flexibility was crucial for our speed to market. So after that, we started analyzing chat by chat, the tractors, understanding what wasn't answering the VA, and what were the people were telling in the VA. So we had to improve our strategy, and we decided to start developing integrations. And last year, in twenty twenty-five, we decided to take a leap and switch to a generative AI engine. And that was a very, very, very important challenge to us because we had to migrate all the incorporated knowledge base. We had to change some flows, and we had to create the VA's personality prompt. So with this big step, we could continue working. And today, I manage a specialized team of ten people that are exclusively dedicated to AI. And currently, our VAI, it handles, like I said before, one hundred fifty-five sessions per month. And it has a retention rate of eighty-eight percent that has transformed our customers' experience very, very grateful, efficient. Exciting.

Chapter Regulatory Challenges in AI Implementation

Dante Franco

Incredible numbers. So let’s talk a little bit about regulation. Regulators in Argentina are watching really closely. How has operating under that regulatory spotlight shaped what you can and can't do with generative AI?

Jesica Cardozo

Well, operating in a regulated, in a heavily regulated market like Argentina means that security is non-negotiable. We work closely with our teams to ensure that while our VA has a generative AI, it is also compliant with, regulations because we are a public bank. And in that context, first, we have worked on security, on security identity validation methods because it’s essential to protect our customers' data. Second, we have to daily manage the VA's personality prompt to make sure that it's not going to hallucinate or request for personal information. And, also, being the first bank in Argentina with generative AI, we are receiving many requests for documentation. So we are, we have proven that we are secure, but maybe this year soon, are going to be audited. So we are, we are waiting for that.

Dante Franco

I'm gonna circle back a little bit. You’ve said that customers don’t always want to be talked to. Don’t always want to talk to a virtual assistant. How do you turn that resistance around, and what role did building out integrations play?

Chapter Customer Trust and Integration Success

Jesica Cardozo

I always say that people don’t dislike AI. People dislike poor customer experiences.

Dante Franco

And sure.

Jesica Cardozo

Our customers are used to talk with human agents. And the culture shift is not easy, but we are focused on achieving a digital transformation that allows us to progress, like I say, beyond today. And, of course, we want to reduce costs. So our strategy is to humanize technology, and we want to create fluid interactions with AI by using all its capabilities. We are looking for a virtual assistant that is empathetic, that understands the context of the interactions. And, of course, we want to solve the customer queries. In that way, our integrations, like I said before, we have credit card balance, purchase authorization. We have a home banking key, and they have played an important role allowing people to get personal information with just a few clicks and without needing to go to the bank. So, for example, this specific integration where we can get the home banking key instead of going to an ATM is the most successful transaction use case that we have.

Dante Franco

So you have now eight integrations live. What was the moment that customers started trusting the bot with more than simple queries?

Jesica Cardozo

Customers started trusting the VA two years ago when they started asking for their credit card balance, and they realized that they could get personal and transaction information by using the virtual assistant. So in that way, people started using the WhatsApp channel, and that has resulted in huge savings, human cost.

Dante Franco

So now after building integrations, you're pushing beyond simple Q&A. What's the next integration that will be a game changer for your customer?

Jesica Cardozo

I’m convinced that the new feature that will change the customer experience completely is making money transfers using WhatsApp.

Chapter Future Innovations: Money Transfers via WhatsApp

Jesica Cardozo

People will no longer need to use the home banking app. This is going to be great for Argentina because we are used to sending WhatsApp messages all day long. So it is the best channel that Argentinians can use to get personal information.

Dante Franco

So now looking forward, where is this all heading, if we’re back here at AASummit next year, what will artificial intelligence agents be doing inside your bank that they are not doing today?

Jesica Cardozo

We are planning to identify customers from the first interaction to customize their experience. And we are currently working on technical details in that moment.

Dante Franco

K. And will AI agents replace call centers within five years? Yes, no, and why?

Chapter The Role of AI in Call Centers

Jesica Cardozo

I think it’s not going to replace them, but I think it will reduce the operating cost of the banks. And in that moment, I think AI agents will have become more empathetic, so people will feel as if they were talking with a human agent. So I think it’s going to help a lot. Not replace them, but it helped a lot.

Dante Franco

So what would you need from regulators or from your technological partners to move twice as fast as you're doing now?

Jesica Cardozo

I think we need more regulations, which are better adapted to the emerging technologies. With our technological partnerships, we need to continue developing integrations. And, of course, we need to understand a little bit more about generative AI to use all its capabilities, for example, in ad channels like Voicebot.

Dante Franco

So what’s one piece of advice you’d give to a bank that hasn’t deployed yet AI agents and is wondering where to start?

Chapter Advice for Banks Considering AI Deployment

Jesica Cardozo

I would love to tell them that it’s not as easy like I thought, like we thought. We have learned that technology is nothing without the customer-centric strategy. And, also, we have learned to listen to our customers' feedback, and that is so, so, so important. The technological partnership is as important as the commercial strategy. So you have to choose someone that could stay with you and give you as much support as you need.

Dante Franco

Super. I’d like to thank you very much. As we start the closing, I have just two more questions. The first one is from your own personal experience, what are the main benefits that artificial intelligence adoption is generating for Banco Colo?

Chapter Benefits of AI Adoption at Banco Colo

Jesica Cardozo

We have three. I can tell you three benefits. Of course, we have lower costs than last year. We reduced human agent attention because of the retention rate of the virtual assistant. That's number one. Number two, the customer experience is so much better than last year. The virtual assistant understands much more the interactions, the context, and the people feel better with this VA. And third, an important tool, my team for example, my team needs less time to work or less time to put in information in the knowledge base. So they have more time to focus on the customer experience that we want.

Chapter Lessons Learned from Implementing Conversational AI

Dante Franco

And after your experience implementing conversational AI solutions for the customer service channel, what lessons learned or practical recommendations would you share with all the financial institutions that are just beginning this journey?

Jesica Cardozo

Also three recommendations. The first one, most people, at least there in Argentina, most people is fear of AI agents. And I would tell that you have to change your mind and start thinking about the VA like a complement, like, complement of your daily work because it will help you in reducing your time work and your operating work. And second, you will have to daily manage with the personality prompt, the VA. You will create it once, but then you have to analyze it every day to make sure that it’s not going to hallucinate, for example, or request for personal information. And the last but not least, don’t forget about the transactional use cases because if you are going to migrate to Azure IoT BI, you have to do some changes on the flows, and you have to take care with that.

Jesse and Dante, thank you so much. Appreciate your time and great example of a successful implementation, how to scale it and watch it grow.