This is the mission that gave rise to MyDatAgent

During Web Summit Rio 2026, we had the honor of being selected by the official internationalization program promoted by ApexBrasil and Sebrae, an initiative that brings together Brazilian startups with global potential and the capacity to represent national innovation in the international market.

Beyond our participation at the event, we were featured in an official Sebrae Nacional article, recognizing our work in developing corporate artificial intelligence focused on digital sovereignty, privacy, and governance.

The problem the market still hasn't solved

The popularization of generative models accelerated AI adoption within companies. But along with this revolution came critical issues that most organizations have not yet addressed:

  • Strategic data being sent to third-party platforms
  • Lack of governance over autonomous agents
  • Unpredictable costs with proprietary model billing
  • Excessive dependence on external providers
  • No control over sensitive corporate knowledge

Many organizations are adopting AI without realizing they are creating new technological dependencies.

Our answer: MDA LLM

Definition
MDA LLM · the private enterprise LLM
An infrastructure that enables organizations to operate their own layer of artificial intelligence in a private, secure, and governed manner. Companies build intelligent agents, copilots, autonomous workflows, and multi-agent systems without relying entirely on external platforms.

The architecture was designed for enterprise environments and includes:

Private LLM

32B parameter model with 256k context, running on dedicated infrastructure — never shared with third parties.

Agent Governance

Full control over what each agent can do, with granular permissions and complete audit trails.

Access Control

Per-user permissions, integrated SSO, and automatic sensitive data masking.

Observability

Complete logs for every interaction, with per-user traceability and real-time auditing.

Dedicated Deploy

Isolated environment per customer, with dedicated VPC and data centers in Brazil.

Enterprise Multi-Tenancy

Complete isolation between departments or branches, with per-tenant policies.

Performance with economic efficiency

Artificial intelligence cannot be just powerful. It needs to be economically viable at scale. With our proprietary optimization layer, we achieve concrete results in production:

-70%
Reduction in AI billing costs
32B
MDA LLM parameters
256K
Context window

This allows companies to increase AI adoption while maintaining financial predictability and improving return on investment.

"More intelligence. Less computational waste. More operational control."

— The principle guiding MDA LLM's architecture

The future of AI will be sovereign

We believe the next generation of enterprise artificial intelligence will not be defined solely by who has the most advanced models. It will be defined by those who can combine:

Privacy

Data that never leaves the company's environment. Zero risk of leakage to third-party platforms.

Governance

Full control over agents, permissions, auditing, and LGPD compliance.

Performance

32B parameters with chain-of-thought reasoning. Complex answers with enterprise precision.

Cost Efficiency

Up to 70% reduction in AI billing. Financial predictability for scaled adoption.

The recognition received at Web Summit Rio and by Sebrae Nacional reinforces our conviction that Brazil can lead the construction of this new global artificial intelligence infrastructure.

"We are building much more than a platform. We are building the infrastructure for the next generation of enterprise AI."

— Fernando Outa, Founder

The future of Enterprise AI is Sovereign.

Executive summary

Key takeaways

  • MyDatAgent was selected by ApexBrasil and Sebrae to represent Brazilian innovation at Web Summit Rio 2026.
  • MDA LLM solves the core problem: companies can have powerful AI without sacrificing privacy, governance, or financial control.
  • Up to 70% reduction in AI billing costs enables scaled adoption with financial predictability.
  • The future of enterprise AI is sovereign: privacy + governance + performance + efficiency.