Brazil is the next ServiceNow release train after Australia. The useful story is not a list of marketing claims: it is how the release may change app delivery, automated testing, AI-agent engineering and source-controlled deployment. This guide separates documented preview capability from roadmap statements and turns the changes into a controlled sub-production test plan.
| Stage | Primary check | Evidence |
|---|---|---|
| Baseline | Australia behavior | Regression record |
| Preview | Brazil capability | Test results |
| Decision | Risk and value | Go / no-go |
| Adopt | Controlled rollout | Monitoring |
Release status and dates
ServiceNow's public version schedule lists Brazil as the Q4 2026 family, with Early Availability beginning in September 2026. ServiceNow community guidance also states that Brazil Personal Developer Instances are expected from 24 September 2026. This is therefore a preview and preparation article—not a claim that every feature is already generally available in every customer instance.
Patch availability, Store application versions, licensing and exact production dates can differ. Confirm the official release notes, Upgrade Center results and entitlements for your own instance before planning a rollout.
| Milestone | Publicly documented timing | What you should do |
|---|---|---|
| Early Availability | September 2026 | Validate an eligible sub-production target |
| Brazil PDI preview | 24 September 2026 | Reproduce critical developer workflows |
| Production planning | Customer-specific | Use Upgrade Center and official release notes |
Dates and preview capability can change. Official release documentation remains the production source of truth.
Developer-platform changes worth testing
Brazil preview material highlights a broader Fluent and Now SDK surface, including APIs for REST messages, aliases, alias templates, retry policies and data lookups. It also describes CLI capabilities for CI/CD, instance queries and explanations. These are useful only when teams prove scope, authentication, update behavior and rollback in their own pipeline.
- 01Create a Brazil PDI or approved Early Availability sub-production instance.
- 02Record the current Australia behavior and baseline evidence before upgrading.
- 03Select one small scoped application with source control and automated tests.
- 04Exercise SDK build, install, query and test operations using least-privilege credentials.
- 05Compare generated artifacts and application files, not only the user-interface result.
- 06Document unsupported files, manual gates, rollback steps and the accountable owner.
| Area | Brazil preview direction | Validation question |
|---|---|---|
| Now SDK / Fluent | Additional APIs and CLI workflow | Can the same source build, install and test across development and test? |
| ServiceNow Studio | Vibe Coding mode, source browsing and Source Control V2 direction | Which artifacts round-trip safely? |
| Flow debugging | Modify inputs during debugging | Is the test isolated and repeatable? |
| Build Agent | Custom rules, skills and assisted build tasks | Where must a human approve generated work? |
AI Agent Studio: a connected engineering canvas
Preview guidance describes a redesigned AI Agent Studio that brings agents, tools and workflows into one connected canvas. It also improves access to testing and evaluation through AI Agent Advisor, versions the wider agent or workflow definition, and can generate an update set for promotion.
A crucial implementation detail is that generated update sets may not include every tool dependency. Inspect the package and dependency list, and promote it through the same controlled path as other platform configuration.
| Test | Expected evidence |
|---|---|
| Version change | Previous and current definitions can be compared |
| Tool dependency | Every referenced tool exists and is authorized |
| Evaluation | Known-good, unsafe and ambiguous prompts have scores and traces |
| Promotion | Preview resolves every dependency without unexpected collision |
Test Agent and Auto Maintenance
ServiceNow preview articles describe Test Agent as an assistant that can author, execute and troubleshoot tests, while Auto Maintenance detects application drift and proposes changes to applicable tests. The value is faster feedback; the risk is trusting generated coverage without checking whether it proves the business outcome.
- Start with one Fluent application and written acceptance criteria.
- Review every generated test step, input and assertion.
- Run positive, negative, role-security and regression cases.
- Require evidence when a test is updated or deleted after application drift.
- Retain manual controls where automation cannot prove the outcome.
Preview guidance says Fluent applications are the strongest fit. Native applications may require more descriptive prompting and manual review.
Autonomous engineering still needs governed delivery
Brazil-era Build Agent material shows an autonomous engineering flow that can read requirements, ask clarification questions, work against an instance, generate tests and package work. Treat it as an engineering contributor, not an unreviewed production deployer.
Require an approved requirement, named owner, least-privilege connection, peer and security review, automated evidence and a human production gate. Availability may depend on AI Workflow Factory, Store versions, subscription or PDI eligibility.
| Control | Minimum evidence |
|---|---|
| Requirement | Acceptance criteria and unresolved assumptions |
| Generated change | Readable diff and owning application scope |
| Security | ACL, role, credential and cross-scope review |
| Quality | ATF plus integration and regression tests |
| Deployment | Preview, approval, rollback and monitoring |
Australia to Brazil: practical comparison
Australia strengthened the enterprise AI and governance conversation. Brazil preview material continues that direction but places more emphasis on the developer delivery loop: source-driven applications, assisted engineering, test creation and maintenance, and a connected agent-building experience.
| Planning lens | Australia-era focus | Brazil preview focus |
|---|---|---|
| AI adoption | Scale useful AI with governance | Build, evaluate and maintain agentic solutions |
| Developer experience | Modern application-development foundation | Expanded SDK, Studio and assisted workflows |
| Quality | Evaluate AI and platform outcomes | Generate and maintain more tests with review |
| Upgrade decision | Adopt proven capabilities | Validate preview features and dependencies |
This is an implementation-oriented comparison of release direction, not a complete feature-by-feature release-note diff.
Two-week Brazil validation plan
- Do not use production data in a preview experiment unless explicitly approved and protected.
- Retest integrations, MID Server paths, scheduled jobs, email and role-based access.
- Compare performance and error rates before and after upgrade.
- Label each item as shipped, Store-delivered, licensed, preview-only or roadmap.
| Day | Activity | Exit evidence |
|---|---|---|
| 1–2 | Inventory custom apps, integrations, agents and tests | Owners and risk-ranked scope |
| 3–4 | Prepare approved sub-production target | Baseline and rollback confirmed |
| 5–7 | Run upgrade scans and critical regression | Failures and owners |
| 8–9 | Prototype one SDK/Fluent delivery path | Repeatable build record |
| 10–11 | Evaluate one agent workflow | Scores, traces and permissions |
| 12–13 | Review generated test maintenance | Approved assertions |
| 14 | Go/no-go review | Decision, exceptions and monitoring |
Official sources used
This is original training content. It paraphrases public ServiceNow material, uses an original diagram, and does not reproduce product screenshots or release-note text.
Explore more implementation-focused ServiceNow and architecture guides.