FEBRABAN TECH 2026 is the main technology and innovation event of the Brazilian Federation of Banks, held between August 24 and 26, 2026 at Distrito Anhembi, in São Paulo. Under the theme "Intelligent Agents, Human Leadership," this edition placed the adoption of AI at the center of the financial sector's agenda, focusing on governance, trust, and control.
What was FEBRABAN TECH 2026?
The 2026 edition occupied an expanded structure of 42,000 m², 71% larger than that of 2025, with six simultaneous stages and programming from 10:00 AM to 6:30 PM over the three days. The event took place at Avenida Olavo Fontoura, 1209, Santana, São Paulo, near the Portuguesa-Tietê station.
What marked this edition?
FEBRABAN TECH 2026 confirmed a shift in the debate on AI in the financial sector: the focus has moved from the adoption of models to operational control over autonomous agents. Discussions revolved around questions that now guide the governance agenda of banks and technology companies:
- How to control operational autonomous agents?
- Who manages the life cycle of these agents?
- What systems, data, and APIs do they use?
- How to identify the associated risks?
- How do you measure costs and consumption?
- How to control identity and access?
- How do you connect agents to business services?
AI is no longer an experiment.
Discussions at the event reinforced the idea that a new class of digital assets is already operating within banks: AI models, agents, co-pilots, APIs, MCP servers, vector databases, integrations, prompts, credentials, accessed data, and the infrastructure that supports it all.
The consequence discussed was straightforward: governing AI is different from using it. Financial institutions face the challenge of discovering which agents exist, who created them, what data they use, which applications they access, and what the cost and risk of each is. This scenario brought the debate closer to already mature IT disciplines, such as IT Asset Management (ITAM), Configuration Management (CMDB), and Cyber Asset Management.
From asset management to AI Asset Management
One of the topics discussed at the event was the extension of ITAM and CMDB practices to new AI assets.
| Discipline | Questions that you answer |
|---|---|
| ITAM | What asset exists, who is responsible for it, where is it located, what is its cost, what is the associated contract, and what stage of its life cycle is it in?. |
| CMDB | What relationships does the asset have with applications, what services does it depend on, what is the impact of downtime, and what technologies is it connected to?. |
| Expansion to AI | Which agents operate, which model does each one use, which system or service do they support, what data do they consult, which tools do they execute, what identity is associated with them, and what risk do they represent?. |
Cyber Asset Intelligence gained ground.
The program also addressed the complexity of financial environments, which combine servers, endpoints, cloud, applications, networks, APIs, connected devices, industrial environments, third-party components, SaaS, identities, and now, AI agents and systems.
The common thread among the presentations was that security depends on visibility: without knowing what assets actually exist, it is impossible to protect them. Unknown, unmanaged, or incorrectly classified assets remain the main blind spots for institutions.
CMDB and trusted data as infrastructure for AI.
Another consensus that emerged from the discussions: AI agents perform tasks quickly, but the quality of decisions depends on the corporate context — applications, services, infrastructure, users, assets, contracts, dependencies, criticality, risks, incidents, and changes.
A mature CMDB establishes relationships between technology and business services, providing the necessary context for automated decisions with control and traceability. The question that dominated the debates shifted from simply "how to implement AI?" to including "on what data does this AI make decisions?".
Main debates
The opening panel, on August 24, brought together executives from Banco do Brasil, Bradesco, Nubank, Santander Brasil, BTG Pactual, Caixa, and Itaú Unibanco to discuss "Intelligent agents, human leadership: how banks are redesigning scale, trust, and value in the financial sector.".
International experts in focus.
The keynote program featured three confirmed international speakers:
- Joel Mokyr, economic historian and 2025 Nobel laureate, with the keynote address “The past and future of innovation: is it possible to sustain progress?”
- Radia Perlman, engineer and Fellow at Dell Technologies
- John Maeda, a leading figure in design and technology, with the keynote address “Networked AI Agents: Is it the end of the Assistant Era?”
Current technological ecosystem
The event brought together exhibitors such as AWS, Google Cloud, Microsoft, IBM, Oracle, CrowdStrike, Dynatrace, Tanium, Salesforce, ManageEngine, Snowflake, Red Hat, Dell, Deloitte, KPMG, PwC, and specialized fintech companies, reflecting how financial innovation today integrates platform, data, infrastructure, security, observability, and automation with governance.
4MATT at FEBRABAN TECH 2026
The 4MATT, ServiceNow Elite Partner in Brazil, participated in the event at booth A3, presenting topics such as AI Asset Management, ServiceNow AI Control Tower, and Cyber Asset Intelligence., ITAM, CMDB, AI asset governance and visibility into risks and technology dependencies.
4MATT's presence at the event aimed to connect the discussion about AI to a concrete operational problem: maintaining visibility, context, control, and governance in the face of the continuous emergence of new digital assets.
Five questions left for technology executives.
The discussions in this edition highlighted five key points for those leading technology in financial institutions:
1. Agent autonomy. The greater the autonomy, the greater the need for controls over identity, authorization, data, and logs.
2. Discovery of AI assets. Decentralized adoption creates agents without knowledge of architecture, security, or governance—discovery and inventory remain fundamental requirements.
3. Responsibility for governance. Security, infrastructure, architecture, data, compliance, and ITAM have different responsibilities; clearly defining ownership remains a challenge.
4. CMDB and operational data for AI. Enterprise agents need context — reliable configuration data increases their ability to support operational decisions.
5. Measuring value. The number of agents or models are adoption metrics, not outcome metrics. Governance connects technology to indicators such as productivity, risk reduction, uptime, quality, cost, and experience.
FEBRABAN TECH 2026 as a barometer for the sector.
The event confirmed that AI has moved beyond the experimental stage and become an integrated part of the processes and operations of the financial sector. The next step discussed at the event requires combining speed with security, financial control, data governance, architecture, and risk management—with disciplines such as ITAM, CMDB, CSDM, Cyber Asset Intelligence, and operational governance gaining critical importance in this movement.
Frequently asked questions about FEBRABAN TECH 2026
When did FEBRABAN TECH 2026 take place? Between August 24th and 26th, 2026.
Where was it held? In the Anhembi District, Avenida Olavo Fontoura, 1209, Santana, São Paulo, Postal Code 02001-900.
What was the theme of the edition? "Intelligent agents, human leadership".
What were the main topics discussed? AI, AI agents, financial innovation, cybersecurity, data, cloud, infrastructure, payments, and customer experience.
Who participated? Executives from banks, technology companies, fintechs, specialists, and consultancies.
Did 4MATT participate? Yes, with booth A3 and presentations on AI Asset Management, ServiceNow AI Control Tower, Cyber Asset Intelligence, ITAM, CMDB, and governance.
Why is governance central to the financial sector? Financial institutions operate in regulated environments with stringent controls over identity, data, security, risk, and continuity. Autonomous AI agents require extending those same controls.