#

Key takeaways

• Itaú is expanding its customer-facing ia.i conversational interface.

• The bank says ia.i reached about 300,000 customers in its initial phase.

• IARA is the internal governance platform behind the rollout.

• The bank reports legal-work automation at large scale and efficiency gains.

• The source does not establish service availability in Morocco.

Itaú's customer AI rollout

OpenAI's October 16 case study describes how Brazilian bank Itaú is expanding two linked systems. The first is ia.i, a customer-facing conversational interface. The second is IARA, an internal governance platform that supports the bank's AI use.

Itaú says ia.i reached about 300,000 customers in its initial phase. The interface can be activated by text or voice. It brings account, credit, card, insurance, and other interactions into one environment. When needed, it also provides a route to human support.

The bank's stated goal is to make the experience available to its entire customer base by the end of 2026. That is a target, not a completed rollout. The source presents this as a planned expansion, not a finished deployment.

What IARA does

Behind the customer interface, IARA provides controlled access to several model providers. OpenAI is one provider in that environment. The platform also includes observability, token-cost oversight, and responsible-AI policies.

These functions matter because they shape how the bank manages AI use. Controlled access can help centralize model selection. Observability can help teams monitor system behavior. Token-cost oversight can help track usage. Responsible-AI policies can guide how the tools are used.

The source does not give technical implementation details beyond those points. It does not describe the full architecture, internal approval process, or any external audit results. It only states the capabilities that Itaú and OpenAI describe in the case study.

Legal-department automation and reported results

The case study also describes a legal-department solution. It operates over a repository exceeding 90 million documents. It runs more than two million tasks per month. Itaú and OpenAI report efficiency improvements of 20 to 80 percent in stages of its processes.

Those figures are presented as reported metrics from the bank and OpenAI. They are not independently verified outcomes across all departments. The source therefore supports the numbers as reported results, not as universal performance claims.

This part of the case study shows a different use of AI from the customer interface. One system serves customers directly. Another supports internal legal work. Together, they show how the bank is applying AI across both external and internal workflows.

Customer DNA and planned future functions

Itaú also describes Customer DNA, a financial-context layer serving roughly 70 million customers. GPT models are among its components. The source does not define every function of this layer, but it places Customer DNA within the bank's broader AI stack.

The case study says future ia.i functions are planned. These include product simulations, sign-ups, and recurring financial analyses. These are described as planned capabilities, not current features. The source does not say when each function will launch.

Governance and operational considerations

The main operational theme in the source is control. Itaú pairs a customer interface with a governance layer. That structure suggests the bank is treating AI deployment as both a product and a managed internal system.

The source also shows that scale matters. The customer interface reached a large initial audience. The legal solution handles a very large document base and a high monthly task volume. Those details indicate that the bank is using AI in environments where oversight and consistency are important.

Morocco relevance

The source reports no Morocco-specific availability or local deployment. For readers, the general lesson is that AI rollouts work best when customer tools and governance controls move together.

Bottom line

Itaú's case study presents a two-layer AI strategy. ia.i handles customer interactions, while IARA manages model access and oversight. The bank also reports large-scale internal automation and plans for more customer features. The source frames these as ongoing and planned efforts, not final outcomes.