I’m now on my second blog post about AI in two months for Auto Rental News. I’m realizing how AI is moving faster into the car rental sector than most of us anticipated.

It will be topic A as we enter planning season for sessions at the International Car Rental Show near Dallas, May 13-15, 2026.

I recently spoke with Julian Gritsch, founder and CEO of JGCorp near São Paulo, Brazil, who has appeared on AI and tech-related panels, bringing a wealth of rental car operator experience and analysis to the subject of AI.

He now welcomes a new business partner, Jul-IA. This multilingual AI agent doesn’t just talk to customers; it works inside the business, taking real responsibilities across sales, finance, and operations, all with a friendly, human-like language.

Since WhatsApp is the dominant text communication channel in Brazil, JGCorp. decided to make it the foundation for Jul-IA, among the first industry AI agents specializing in car rental and mobility.

A Modeled Employee

The creators of Jul-IA programmed it to join the organizational chart of a rental car company. It offers these key functions:

• Pre-sales and reservations, 24/7

• Post-sales support, customer service, and FAQs

• Billing and collections via card integrations

• Risk validation through local rental risk platforms

• Traffic fines management and compliance workflows

• Deep integrations with CRM, ERP, and even global brands like Europcar Brazil’s booking system

Jul-IA officially launched in September at a Brazilian car rental event similar to the International Car Rental Show, in Foz do Iguaçu, Brazil. Jul-IA acted as the event’s virtual host, handling registrations, sending schedules, connecting sponsors with attendees, and even giving weather updates. Jul-IA live can be found at www.jgcorp.com.br/julia

“Our goal is not to build 'another chatbot,' but to create a specialized intelligence that truly understands how a rental business operates — speaking the same language, using real operational logic, and continuously learning from user interactions,” Gritsch said.

Explaining Jul-IA’s 'Make-Up' (Configuration)

Jul-IA is a proprietary AI system designed specifically for the vehicle rental industry. As Gritsch explains, Jul-IA doesn’t rely only on a general large language model (LLM); it’s powered by a retrieval-augmented generation (RAG) structure, which draws on and connects directly to JGCorp’s 25 years of practical experience in and guidelines for car rental operations.

In sum, Jul-IA combines three layers:

1. A general LLM, like the foundation used in models such as GPT or Claude.

2. A specialized RAG built from my technical materials, frameworks, and case studies in the rental sector.

3. A layer of intelligent agents that act based on real business workflows — from managing fines and maintenance to customer communication and pricing.

During testing, the company compared the three model configurations. The RAG layer already made the answers far more accurate and context-aware, but when combined with the agentic layer, Jul-IA was able to reason and act, Gritsch said. 

For example, it automatically sent customer updates, analyzed fleet data, and cross-referenced contract terms.

How Jul-IA Handles Their Workdays

Gritsch and I chatted further about Jul-IA’s workdays, without AI assistance or invasive bots:

Q: Interesting name, Jul-IA. What’s the origin of that?

A: It’s a mix of my own name, Julian, and IA, which stands for Inteligência Artificial in Portuguese. Unlike English, where we say AI, in Brazil we say IA. So “Jul” plus “IA” became Jul-IA.

Q: How has this AI project evolved?

A: It’s very different from earlier this year. Earlier this year, we were still teaching rental companies how to use AI. Now, we’ve developed a real product—an intelligent virtual agent we call Julia. About eight to ten companies are already using it, including large ones like Europcar in Brazil. We just signed a deal with Stellantis, formerly Fiat, as well.

Q: What exactly does Jul-IA do for rental companies on a daily basis?

A: Jul-IA is an AI assistant that handles the full rental journey, from the first customer inquiry to after the rental ends. In Brazil, most customer interactions happen on WhatsApp, so Jul-IA works seamlessly there. When a customer starts chatting, say, asking about available cars, Jul-IA provides the answers automatically. What sets Jul-IA apart is that it’s powered by a large database we built specifically for the car rental industry. It’s based on an e-book I created years ago -- a detailed guide to Brazil’s car rental business. So, when a company installs Jul-IA, it already comes loaded with essential industry knowledge and terminology. The company only needs to customize the agent for its specific business model, such as rentals for app drivers or long-term corporate clients.

Q: So, the tool already “knows” how the car rental business works?

A: Exactly. It’s pre-trained with that foundation, and we keep improving it with more data and better responses over time.

Q: How does Jul-IA integrate with rental company systems?

A: Jul-IA connects directly to a company’s CRM, like Pipefy or others. [Pipefy is a global artificial intelligence platform that orchestrates autonomous AI agents]. 

When a customer interacts with Jul-IA, all their details are automatically entered into the CRM: Name, contact, rental stage, and so on. Even if a customer stops mid-process, the company still retains their data for follow-up. In cases where there’s no API access, Jul-IA can simulate a human user. For example, at Europcar Brazil, the company’s API wasn’t available, so Jul-IA retrieves data directly from their website, filling in forms and fetching prices automatically.

Q: That’s clever. And what about payments?

A: In Brazil, besides credit cards, many people use Pix, a digital instant payment system like Venmo. Jul-IA connects with Pix APIs, allowing customers to pay directly through the chat.

Q: Can Jul-IA verify customers’ identities or credit backgrounds?

A: Yes, absolutely. Brazil uses a system called Serasa, like Experian or Equifax in the U.S. Jul-IA can automatically check a customer’s credit score using their CPF number, which is like a Social Security number. It can also check driving licenses through the national database, such as if a driver has too many points. Jul-IA flags them as ineligible to rent.

Q: That’s impressive. What happens if a customer asks a question that Jul-IA isn’t trained to answer, such as something about car seat height or accessibility features?

A: Jul-IA politely explains that it doesn’t have that information and refers the customer to the company’s reservation or help desk. Some companies even have Jul-IA forward the question to their call center, which then follows up directly. We also program Jul-IA to steer conversations back to car rentals if someone tries to go off-topic.

Q: Can Jul-IA handle travel-related questions, such as restaurant or hotel recommendations?

A: Surprisingly, yes. One customer asked for restaurant suggestions in São Paulo, and Julia gave two great options. It then even recommended which one was better! It can also check weather forecasts for destinations. With its RAG, it pulls about 85% of its responses from verified internal data and 15% from the web.

Q: Does Jul-IA learn from each interaction?

A: Yes. The more customers use it, the smarter it becomes.

Q: Have you noticed patterns among users, such as younger people being more comfortable with it?

A: Definitely. Younger customers prefer contactless, self-service experiences. In Brazil, everyone uses WhatsApp, even people over 50, but younger generations almost never want to call anyone. In fact, telemarketing is so overused here that most people ignore phone calls. That’s why text-based AI services are so successful.

Q: And for older customers (like me) who still prefer to talk to a person?

A: Jul-IA can transfer them to a live call center anytime.

I informed Gritsch that I would like the most updated version of Jul-IA to debut at ICRS next spring, allowing us to assess how they handle a live audience reserving rental cars. Maybe by that time, Jul-IA could moderate the session, too.