ServiceNow Brazil: AI that finishes the work

From starting work to finishing it

Most enterprise AI still helps with a step: it drafts the reply, summarises the ticket, suggests the fix. A person then carries the work to the finish line. ServiceNow’s Brazil release aims to change that split, with AI agents that complete whole pieces of work while people stay in charge of the outcome.

The question for every organisation is no longer whether AI can do the task. It is how much work you are prepared to hand over, and what keeps you in control when you do.

What Brazil is

Brazil’s tagline sums up the intent: “One AI experience. Autonomous work by design.” ServiceNow is moving AI from doing tasks to doing whole pieces of work, with governance built in rather than bolted on.

The release pushes AI into three frontiers, each with one headline capability:

• Software development (“Build with AI”): AI that builds and tests, not just suggests code.
• The desktop (“Take AI further”): AI that works across the applications and files people use every day.
• Enterprise systems (“Connect every agent”): AI agents, including third-party ones, acting inside governed workflows.

Three capabilities, one pattern

Each capability gives AI more room to act, and each names who keeps control.

Autonomous Engineer takes on a full unit of development work rather than a single suggestion. Developers move from writing every line to setting direction and judging what is ready for production.

Arc changes how employees use AI day to day. Instead of walking an assistant through each step, they delegate the outcome they want and let Arc work across the tools involved.

Action Fabric opens the platform to the wider agent ecosystem. Agents from other vendors can act inside ServiceNow workflows, under the same rules as everything else on the platform.

Governance is the accelerator, not the brake

More autonomy raises an obvious concern: who is accountable when an agent acts? ServiceNow’s answer is that people stay informed, guide what happens next, and remain in control.

The argument is that governance is what makes autonomy scalable. Because identity, permissions, policy and workflows stay connected on one platform, an agent inherits the same rules a person would. Organisations can then extend AI to more work without building a new control layer for every agent or tool.

In practice, this is the difference between a promising pilot and AI running in production. Without clear permissions and an audit trail, most organisations will never let an agent finish real work.

Where Brazil fits

Brazil is the latest step in a steady progression, from AI agents handling routine work to governed, autonomous work across the enterprise. The release taglines tell the story, newest first:

Brazil also delivers on the themes ServiceNow set out across 2026. At Knowledge 2026 (5–7 May) the message was “Welcome to Agentic Business”, with CreatorCon promising developers they could “Build Anywhere. Run on ServiceNow.” In August, the Super Intelligence Control Tower focused on “governance that moves at machine speed”. Brazil brings both threads together: more autonomy, with control designed in.

Our view: what Brazil means for your organisation

At Sequal Consultancy, we see Brazil as a release to plan for deliberately, not just install. The value depends less on switching features on and more on how ready your platform and your people are to hand over whole pieces of work.

Four questions we recommend asking before your next upgrade:

1. Is your governance foundation solid? Brazil’s control model relies on clean identity, roles, permissions and policies. Gaps there become gaps in what agents can safely do.
2. Which work is ready to delegate? Start with well-defined, repeatable outcomes where success is easy to check, then widen the scope as confidence grows.
3. What agents already exist in your landscape? If teams use AI agents from other vendors, Action Fabric is a chance to bring them under one set of rules rather than leaving them ungoverned.
4. Who owns the outcome? Agree in advance who reviews agent work in development, IT, security and the business, so control is a role, not an afterthought.

Organisations that answer these early will get more from Brazil, and from the releases that follow it.

Next steps

ServiceNow has published a range of material to help teams get started with Brazil: an upgrade kit, the launch replay with demos, release notes, community pages for each role, a Workflow blog post on the three frontiers, and live “Explore what’s new” expert sessions.

Planning your move to Brazil, or wondering where autonomous work could pay off first? Get in touch with Sequal Consultancy and we’ll help you build a roadmap that balances speed with control.