✕
Releases

ServiceNow ITOM: Top 9 Brazil Release Features and Use Cases

By Mahathi Veena

For years, the ITOM conversation has centred on visibility. What infrastructure do we have? What services depend on it? What is healthy, what is failing, and where is the impact? After reading the ServiceNow Brazil release notes, I think the more interesting story is what comes after visibility.

From Discovery and Service Mapping to AIOps, DEX, automation, and MCP, Brazil points towards a more connected model for intelligent IT operations. Here’s my breakdown of Brazil’s top features and enhancements for ITOM. 

The ServiceNow Brazil ITOM Intelligent Operations Loop

Discovery is widening how we identify the estate. Service Mapping adds new ways to work with live service topology. Event Management brings more transparency to AI-assisted alert handling. Health Log Analytics gets better at separating momentary anomalies from conditions that deserve attention. Digital End-User Experience moves from detecting endpoint problems towards fixing them. Model Context Protocol (MCP) is starting to change how authorized users and AI-enabled workflows interact with operational context.

Each is a useful enhancement on its own – together, they look like a connected operating model. 

Discovery and Agent Client Collector establish visibility. CMDB and Service Mapping add dependency and service context. Event Management, Health Log Analytics, Synthetic Monitoring and DEX sense what is happening. AIOps helps interpret what matters. 

Where there is enough confidence in both the condition and the response, remediation and automation start to close the gap between detection and action.

MCP doesn’t fit neatly as a final step. It shows up across several ITOM capabilities, so I see it as an interaction layer across the whole model.

When I step back from the individual enhancements, the operating model I see emerging moves through five increasingly valuable questions: 

  • See: What do we actually have?
  • Understand: What does it support?
  • Sense: What is happening right now?
  • Interpret: Does it matter?
  • Respond: Do we know enough to act safely?

MCP cuts across all five layers rather than sitting at the end, because it changes how people and AI-enabled workflows interact with the ITOM capabilities underneath.

Intelligence sits at the top. The trust starts at the bottom.

Each stage depends on the context beneath it, which is why the story doesn’t start with AI. It starts with one simple question: what is actually there?

1. Discovery: What Actually Exists?

Everything higher up the ITOM stack depends on being able to trust the infrastructure underneath it.

PowerShell 7 support for Discovery arrived with the Australia release. The Brazil Discovery release notes add two MID Server configuration parameters that let File-based Discovery and Windows ADME prefer PowerShell 7 over the default PowerShell 5, on the MID Server and on remote target hosts. 

They also add options for credential-less Discovery: optional UDP port checks, and operating-system guessing where an exact OS match can’t be confirmed. Each credential-less option is enabled through a MID Server property and is inactive by default. Brazil Early Availability also adds daily performance snapshots for slow-pattern tables.

READ MORE: How to Install a ServiceNow MID Server – And Why You Need One

Key Capabilities

  • Configure MID Server parameters to prefer PowerShell 7 for supported File-based Discovery and Windows ADME scenarios.
  • Optionally include UDP ports in credential-less Discovery scans.
  • Allow credential-less Discovery to guess the operating system when an exact match cannot be established.
  • Use daily performance snapshots to gain visibility into slow Discovery patterns as part of Brazil early availability.

Why It Matters

For organizations modernizing their Windows administration, the PowerShell 7 option means Discovery doesn’t have to stay on the default PowerShell 5.

The credential-less options address a familiar problem: large estates always have areas where full credential coverage is hard to achieve. The extra scan options add some visibility and can show where richer, credential-based Discovery is needed.

But broader Discovery doesn’t automatically mean a better CMDB. The question isn’t just “Can ServiceNow discover this?” It’s also:

  • Should this source populate the CI? 
  • How will that CI be identified? 
  • What happens when another source describes the same CI differently?’

At that point, Discovery is part of a wider CMDB data strategy, not just a scanning tool.

Real ITOM Use Case

Imagine an enterprise moving its Windows estate to PowerShell 7 while some network segments still lack full Discovery credentials. File-based Discovery and Windows ADME can be set to prefer PowerShell 7, and the credential-less options can reveal more about infrastructure that is otherwise hard to identify.

Important Considerations

Credential-less visibility is not the same as an authoritative CI population. MID Server architecture, credentials, CI classes, identification and reconciliation rules, and source authority still decide whether the result improves the CMDB.

Discovery can tell us more. We still have to decide what we trust. Once we know what exists, the next question is what depends on it.

READ MORE: ServiceNow Data Foundations (CMDB and CSDM) Certification Guide

2. Service Mapping: What Depends on What?

If Discovery answers “What do I have?”, Service Mapping asks “What does it support?”

Brazil brings two notable Service Mapping developments: MCP tools, delivered through the CMDB MCP Server plugin, and the Lightweight Service Model, available with Service Mapping Plus.

Key Capabilities

  • Query live application-service topology through MCP tools.
  • Use MCP tools to identify mapping gaps.
  • Create application services through supported natural-language interaction.
  • Convert Dynamic and Tag-Based services to the Lightweight Service Model.
  • Maintain current service topology without historical snapshots, reducing storage overhead and improving performance for large CI inventories.

Why It Matters

The value of Service Mapping lies less in the map than in the decisions its relationships make possible. Knowing a database server is unhealthy is far more useful when we also know which application services depend on it.

MCP adds another way to get at that context. Instead of navigating to the topology, an authorized AI-enabled client can query live application-service information in natural language.

Real ITOM Use Case

An infrastructure component is degrading, and the operator needs to know: “Which application services potentially depend on this CI, and where do we have gaps in the current topology?“

Service Mapping MCP tools offer another way to interrogate the live service information behind those questions.

Important Considerations

Making topology easier to access doesn’t fix the topology. Natural-language access makes it easier to consume, not more trustworthy: missing or incorrect relationships stay that way whether someone reaches them through a map, an API, or a conversational interface.

That takes us from infrastructure visibility to service context. The next challenge is deciding which signals deserve attention when that environment starts telling us something is wrong.

3. Event Management: Which Signals Deserve Attention?

Monitoring platforms are extremely good at generating signals. The harder problem is deciding which ones need action.

The Event Management release notes position it around ingesting events, reducing noise through deduplication and filtering, correlating related events into actionable alerts, and connecting alerts to affected services and CIs. The more interesting Brazil developments appear in the associated Service Operations Workspace release notes, particularly around AI-assisted alert operations and monitoring integrations.

Key Capabilities

The AIOps and Express List September Store updates include:

  • Monitoring and supervising agentic workflows through the AI Processing Experience.
  • Visibility into workflow execution states and steps, and access to the reasoning behind them.
  • Operator intervention where required.
  • AI-assisted alert-grouping configuration from the Alert Automation pages.
  • Broader AI Specialist assignment, including an All Groups option.
  • Visibility into the decision, reasoning, supporting evidence, and execution traces behind AI-auto-closed alerts.

The September Integrations Launchpad update also supports conversational connector setup: administrators can configure vCenter, Nagios XI, SolarWinds, SCOM and Zabbix by chatting with ServiceNow Otto. A new Dynatrace Grail connector can create alerts from Dynatrace Grail problem events, with information including the root-cause entity, all affected CIs, a matching severity and the problem’s current lifecycle status.

READ MORE: The Ultimate Guide to ServiceNow Integrations

Why It Matters

The benefit lies in reducing noise so people can focus on the conditions that matter. But there’s a second point too: if AI is going to make operational decisions, operators need to see how those decisions were made.

Closing an irrelevant alert automatically removes work. Being able to inspect why it was closed, which evidence supported the decision, and what happened during execution adds something more important: operational explainability.

Real ITOM Use Case

Imagine an operations team receiving thousands of alerts. An AI-assisted process decides a subset needs no operator intervention and closes them. Instead of the trail ending there, an operator can examine the decision, reasoning, supporting evidence, and execution traces behind the closure. 

As AI gets more autonomy in operational workflows, that transparency matters more.

Important Considerations

Better AI-assisted alert handling doesn’t make up for weak operational context. If an alert is tied to the wrong CI, a service relationship is incomplete, or the monitoring logic is poorly designed, automation won’t fix the underlying problem. It’s most useful when operators don’t have to choose between efficiency and explainability.

There is also a practical migration point: the Dynatrace Grail connector replaces the classic Dynatrace event integration, so teams still on the classic integration will need to move to keep receiving Dynatrace alerts.

But even a perfectly correlated alert raises another question: is the behaviour actually important, or merely abnormal?

READ MORE: Closed-Loop Operations Connect AI-Driven Insights to Actions for ServiceNow Customers

4. Health Log Analytics: When Does Abnormal Mean Important?

This is one of my favourite Brazil enhancements because it addresses a very practical problem: an anomaly is not necessarily an incident. The Health Log Analytics release notes introduce Sustained Alert feedback, which lets operators suppress short-lived anomaly spikes and focus on conditions that persist continuously.

Key Capabilities

  • Suppress alerts generated by short-lived anomaly spikes.
  • Define a fixed duration that determines what counts as sustained.
  • Apply Sustained Alert feedback per metric, from the Alert Card or Express List, individually or in bulk.
  • Inspect the full Smart Parser Gen AI parsing and alert-generation flow, supported by expanded automated testing.
  • Set up MID-less integrations through natural-language prompts using an AI-enabled MCP client when the administrator has ITOM MCP Server Console access.

Why It Matters

Infrastructure and application behaviour is rarely perfectly stable. CPU utilization spikes, batch jobs create predictable demand, and applications see temporary load. 

The useful question is not always “Did something deviate from normal?” Sometimes it is: “Did it stay abnormal long enough for somebody to care?”

Sustained Alert feedback gives teams another way to make that distinction.

Real ITOM Use Case

Imagine a nightly batch process that causes a short but predictable performance spike. The anomaly is real, but if the system returns to normal on its own every time, surfacing it repeatedly is just noise. 

A sustained-duration threshold can separate that from a condition that stays abnormal and genuinely warrants investigation.

Important Considerations

Threshold design still matters. Set the duration too low, and the noise remains – set it too high and meaningful degradation may run longer before it becomes actionable.

Great operations is about understanding which deviations deserve attention. Sometimes, though, what matters about a condition isn’t just that an endpoint is failing. It is how many services depend on it.

5. Synthetic Monitoring: Who Else Is Affected?

Modern application services rarely operate independently. The same authentication endpoint, API, or integration service may support several business-facing applications. A 2026 Synthetic Monitoring Store release improves how those shared dependencies are represented.

Key Capabilities

  • Associate one HTTP endpoint or discovered API with multiple application services.
  • Preserve the CMDB relationships between that endpoint and each service.
  • View those relationships through a read-only Related Services list.
  • Keep the ECC queue clean: the auto-purge job removes old completed Synthetic Monitoring messages and flags stuck or orphaned entries as errors.
  • Show the actual HTTP error response, such as a 401, for MID Server-executed checks instead of reporting “No response”.

Why It Matters

Modern services share dependencies, and monitoring models should reflect that instead of duplicating it. Representing the same endpoint as unrelated monitoring objects just because several applications consume it hides the real architecture.

Real ITOM Use Case

Consider one authentication API consumed by:

  • A customer portal.
  • A mobile application.
  • A partner portal.

One endpoint failure could affect all three. The single monitored endpoint can now keep its relationship with each service that relies on it, so that context isn’t lost.

Important Considerations

The monitoring capability can preserve the relationships, but it still depends on the organization knowing which services actually consume the endpoint. Once again, the underlying service model matters.

Once we know something is wrong and who may be affected, the model reaches its most interesting stage: do we know enough to act safely?

6. Digital End-User Experience: From Detection to Action

There’s a big difference between knowing someone is having a technology problem and being able to do something about it. Digital End-User Experience expands DEX remedial actions for common Windows endpoint conditions and lets those actions run from Flow Designer.

Key Capabilities

  • Start or stop a Windows service.
  • Clear the system cache.
  • Reset the print spooler.
  • Execute remedial actions from Flow Designer.
  • Rate-limit metric and event rules per rule on the device.
  • Use an extension point to gain additional control over how metric-rule events are evaluated before alerts are triggered.

Each remedial action runs only when the device is online and the action applies to it.

Why It Matters

Finding a poor digital experience is useful. Fixing a known, repeatable cause before it becomes another support request is far more useful. This is where the model visibly starts moving from sense towards response.

Real ITOM Use Case

A conventional journey might look like: Poor experience → user reports problem → Service Desk investigates → technician restarts Windows service.

For a sufficiently understood and controlled scenario, that journey can start to look like: Poor experience → condition detected → approved remediation → outcome checked → human intervention where required

The gain is more than fewer clicks, as a known condition can be dealt with before the employee has to start the support journey.

Important Considerations

The goal should be appropriate automation, not maximum automation. Teams still need to understand:

  • The condition triggering the action.
  • The potential blast radius.
  • How successful remediation will be validated.
  • What happens when remediation fails.
  • Where human intervention should remain mandatory.

Just because something can be automated, that doesn’t mean it always should be.

7. Agent Client Collector: What Do We Know About the Endpoint?

Knowing an endpoint exists is one thing. Knowing what software is on it, how it got there, and whether it is actually used gives operations and asset teams a much richer picture. The Brazil-era Agent Client Collector updates, spread across the ACC Framework and ACC for Visibility Content Store releases, expand considerably in that direction.

Key Capabilities

  • Discover portable software installed through package managers.
  • Create custom software-filtering rules to exclude irrelevant software from SAM inventory.
  • Configure licence-key Discovery rules.
  • Categorize discovered software.
  • Improve Windows application usage tracking.
  • Discover and categorize browser extensions and software packages.
  • Protect Windows-agent uninstall through a maintenance token.
  • Support agents operating within non-persistent VDI environments.

Why It Matters

There is an interesting progression here: Device exists → Software exists → How was it installed? → How should it be categorized? → Is it actually being used?

Package-manager discovery matters most in modern engineering environments, where conventional installed-software mechanisms may not give the full picture.

Real ITOM Use Case

Consider a developer estate with traditional Windows applications alongside software installed through package managers, browser extensions, and non-persistent virtual desktops. A conventional software inventory may capture only part of that. The new ACC capabilities add ways to identify and categorize those components.

Important Considerations

Richer collection raises a familiar architecture question: being able to collect an attribute doesn’t make the collector authoritative for it. An organization running ACC alongside Microsoft Intune, MECM/SCCM, Discovery and other endpoint-management sources still needs deliberate identification, reconciliation and source-authority decisions. More data only helps once we know what role it should play.

The same applies to automation – having more capability doesn’t tell us where automation will deliver the most value.

8. LEAP: What Should We Automate Next?

Most operations teams know they should automate more, but the harder question is what to automate first.

The latest LEAP release introduces automation projects across multiple incident taxonomies, which ingest and cluster incident information into unified automation opportunities. It also brings knowledge-base creation improvements and a LEAP MCP tool for external AI clients.

Key Capabilities

  • Analyze incident information across multiple configured taxonomies.
  • Identify incident clusters and potential automation opportunities.
  • Select the target knowledge base and category when drafting a knowledge article from an automation opportunity.
  • Give external AI clients a LEAP MCP tool to check app setup and grouping-pipeline readiness before invoking other tools.

Why It Matters

Automation initiatives don’t always struggle because the technology is missing. Often the harder problem is knowing where automation would deliver real value. If historical incidents form meaningful clusters, operational history becomes a source of automation candidates.

That does not mean AI has decided what to automate. It means the data helps teams ask a better question: where should we look?

Real ITOM Use Case

An organization has a large volume of similar infrastructure incidents, and LEAP analysis identifies a meaningful cluster with potential for automation.

Instead of relying entirely on workshops or individual experience to find the opportunity, historical operational data helps surface it for further analysis.

Important Considerations

Repeated doesn’t necessarily mean deterministic. An activity that happens thousands of times is not automatically a candidate for unattended automation. Exceptions, failure handling, approvals, rollback, and human judgment still need to be understood.

Spotting an automation opportunity should be the start of the design process, not the end of it.

Which brings us to something that sits slightly differently from the rest of the model: MCP.

9. MCP: An Interaction Layer Across Intelligent Operations

One technology keeps appearing as we move through Brazil ITOM: MCP.

The IT Operations Management release notes introduce the ITOM MCP Server Console, installed from the ServiceNow Store, which lets agents retrieve ITOM information and perform supported actions through AI-enabled MCP clients, using natural language outside the conventional ServiceNow UI. It includes AIOps, Service Reliability Management, and LEAP tools – and the pattern extends across ITOM.

Key Capabilities

  • The ITOM MCP Server Console provides an AI-enabled MCP interaction mechanism for ITOM data and supported actions.
  • Service Mapping exposes live application-service topology through CMDB MCP Server tools.
  • Health Log Analytics allows appropriately authorized administrators to configure MID-less integrations through an AI-enabled MCP client.
  • LEAP adds an MCP tool for external AI clients, and LEAP tools are also included in the ITOM MCP Server Console.

Why It Matters

Viewed separately, these are distinct product changes. Side by side, they raise a bigger question: “What happens when operational context stops being something an engineer has to navigate towards and becomes something an authorized AI-enabled client can ask about?”

MCP does not replace the CMDB, Service Mapping, Event Management, or the workflows underneath ITOM. It changes how people and AI-enabled workflows interact with them, which is why I see it across the loop rather than as another stage at the end.

Real ITOM Use Case

Imagine an investigation that traditionally means moving between an alert, its CI, service topology, and related operational information. With MCP, an authorized client could query that context in natural language and invoke supported capabilities where permissions allow. The CMDB and ITOM applications are still there – what changes is how people interact with them.

Important Considerations

Making powerful operational capabilities easier to invoke raises the stakes for:

  • Authentication.
  • Authorization.
  • Read vs. Write permissions.
  • Underlying CMDB and topology quality.
  • Auditability.
  • AI model controls.
  • Human oversight for consequential actions.

AI in IT Operations Management also notes that access to individual AI capabilities can depend on licensing, activation, model-provider availability, and regional restrictions.

Brazil Pattern

MCP interests me more as a cross-ITOM pattern than as a single feature. It may well be a new interaction layer for IT operations.

Bonus Brazil ITOM Capabilities Worth Watching

Not every worthwhile capability needs a full chapter.

Certificate Inventory and Management

Certificate expiry is predictable, yet it can still cause a highly visible service disruption. The Brazil documentation for Certificate Inventory and Management covers discovering and inventorying TLS certificates, along with lifecycle processes including renewal, update, and revocation. 

The opportunity goes beyond keeping an inventory: Discover → establish context and ownership → identify expiry risk → renew → intervene before failure.

Inventory creates visibility – ownership and lifecycle processes turn it into resilience. I treat this as a complementary Brazil capability, not one where every lifecycle function is new in Brazil.

DEX Rule Rate Limiting

DEX metric and event rules can now be rate-limited per rule at the device level, so one noisy rule can’t flood the instance with alerts or events. 

It’s a small change with a bigger lesson: detection stops being useful when the detection mechanism itself becomes a source of noise.

Smart Parser Gen AI

Health Log Analytics now shows more of the Smart Parser Gen AI parsing and alert-generation flow, backed by expanded automated testing. 

The point is not just the Gen AI. AI-assisted operations are far more useful when operators can build confidence in the signals being produced.

Final Thoughts 

What interests me most about Brazil is not any one feature but how the pieces reinforce one another: visibility feeds service context, context gives signals meaning, and signals can increasingly lead to action. MCP adds another way to interact with all of it.

But there is an order to this: build the foundation first, connect the context next, then lay the road for AI. 

Discovery establishes what exists. A well-governed CMDB determines what can be trusted. Service Mapping shows how infrastructure connects to the services the business depends on. Operational telemetry supplies the signals. Only then does AI inherit the context it needs to make useful interpretations and support more sophisticated actions.

Which brings me back to the principle at the centre of the model: Intelligence sits at the top. The trust starts at the base.

AI cannot make an inaccurate CI accurate, it cannot manufacture a missing service relationship, and automation cannot compensate for a condition that was misunderstood. The sophistication of a decision cannot make up for weakness in the context feeding it.

So perhaps the most useful question to take from Brazil isn’t: “How much AI can we enable?” It is: “Have we built an operational foundation that we are prepared to let AI reason over and, eventually, act upon?”

For me, that is the road towards intelligent operations: Discover accurately, model deliberately, observe intelligently, automate selectively, and then give AI something worth trusting.

Further Reading

Want to see more of what’s in the Brazil release? Check out these articles to see more of its new features and updates:

The Author

Mahathi Veena

Mahathi is a Certified Technical Architect and a two-time ServiceNow MVP (2025 and 2026), acknowledged for her technical leadership and community contributions.

Leave a Reply