SolarWinds has introduced Avior, an agentic AI system built to help network teams automate operational tasks, reduce alert noise and improve infrastructure management while preserving administrative control over autonomous actions.

The launch moves SolarWinds beyond AI features that primarily analyze data and offer recommendations. Avior adds goal-oriented agents that can execute tasks across the company's on-premises and SaaS products, using operational data and existing management tools to resolve issues in enterprise environments.

Key Facts

• SolarWinds launched Avior, an agentic AI layer that connects to its monitoring and management products.
• The first offering is a network engineering agent that can automate alert management, configuration automation, compliance checks and incident remediation.
• Avior can correlate alerts into incidents, identify configuration drift, audit devices against policies, manage backups, track capacity, monitor utilization and flag obsolete assets.
• Customers can choose assistive, human-in-the-loop or fully autonomous modes, depending on how much control they want to delegate.
• Safeguards include risk-based guardrails, audit trails, automated rollback, data masking and least-privilege access.
• Avior works with third-party systems such as ServiceNow and is available through a guided onboarding program.
• SolarWinds plans a database administration agent later in Q4, followed by IT administration, SRE and security analyst agents across late Q4 and early 2027.

From Recommendations to Action

Cullen Childress, interim CEO and chief product officer at SolarWinds, said the company has spent four to five years building its AI capabilities. Avior extends that work with agents that can act autonomously and free network engineers to focus on higher-level projects.

The first Avior offering is a network engineering agent. It can automate alert management, configuration automation, compliance checks and incident remediation. SolarWinds chose networking as the starting point because network infrastructure monitoring is a core part of its customer base. Many customers rely on the company's network monitoring portfolio, making network operations a natural entry point for agentic automation.

The network-focused agent can correlate alerts into actionable incidents, identify configuration drift, audit devices against compliance policies and manage configuration backups. It can also track capacity, monitor infrastructure utilization and identify obsolete assets. Those capabilities target common sources of operational friction: too many alerts, undocumented configuration changes, compliance gaps and aging equipment that remains in service without clear ownership.

Alert Noise and Operational Toil

Network teams often face a flood of alerts from routers, switches, firewalls, wireless systems and other infrastructure. Many alerts are repetitive or related to the same underlying issue. Avior's network engineering agent is designed to group related signals into incidents, giving operators a clearer view of what requires attention. The agent can then execute remediation steps or prepare them for approval, depending on the mode selected by the customer.

Configuration drift is another target. When device settings change outside approved processes, teams can spend hours comparing configurations, tracing changes and restoring compliance. Avior can identify drift and audit devices against policies. It can also manage configuration backups, which helps teams recover from unwanted changes or device failures. Capacity tracking and utilization monitoring add a forward-looking layer, helping teams spot constrained infrastructure before performance suffers. The ability to identify obsolete assets can support refresh planning and reduce exposure to unsupported equipment.

Three Modes of Control

SolarWinds designed Avior to give network managers flexibility over how much autonomy agents receive. In assistive mode, agents identify issues and recommend corrective actions, leaving execution to human operators. In human-in-the-loop mode, administrators must approve actions before agents make changes. In fully autonomous mode, agents can complete tasks independently within customer-defined controls and report results back to network administrators.

Childress said customers are unlikely to let agents act without feeling in control, and different organizations have different levels of conservatism. That range is reflected in the control model. A network manager might allow an agent to automatically address repetitive alerts to reduce noise but require approval before a configuration change. An enterprise might permit autonomous storage optimization in staging environments while maintaining tighter control in production. The approach lets teams delegate low-risk work first and expand autonomy as confidence grows.

Guardrails and Safeguards

Avior includes several safeguards intended to keep autonomous activity within bounds. Risk-based guardrails require human approval for potentially dangerous actions. Audit trails track agent activity, decisions and changes, providing a record for troubleshooting and governance. Automated rollback supports reversing changes when necessary. Data protection features mask customer data and restrict access based on user permissions. Least-privilege access limits agents to the tools and permissions needed to complete assigned tasks.

Those controls matter because agentic systems can take actions that affect production infrastructure. An agent that can change configurations or remediate incidents needs clear boundaries, traceability and a way to undo mistakes. Risk-based approvals allow organizations to define which actions are safe for automation and which require human judgment. Audit trails help administrators understand why an agent acted and what changed. Rollback provides a safety net when a change produces unexpected results. Data masking and permission-based access reduce the risk of exposing sensitive information. Least-privilege access narrows the blast radius if an agent behaves incorrectly or is manipulated.

How Avior Connects to Existing Tools

Avior is a separate agentic AI layer that connects with SolarWinds products, including SolarWinds Observability. The agents can draw on telemetry, topology, configuration and other operational data available in those products. Administrators manage the agents and their activities through Avior. For an enterprise using SolarWinds Observability, activating Avior's network engineering agent gives the agent access to the tooling and capabilities inside the product, much as a human operator would use them.

The system also works with third-party systems such as ServiceNow to execute workflows. That integration matters because network operations rarely happen in a single platform. Incidents, changes and approvals often move through IT service management tools. By connecting to those workflows, Avior can help close the gap between monitoring, decision-making and action. The goal is to reduce repetitive tactical work and let network engineers concentrate on strategic initiatives.

Augmenting Teams, Not Replacing Them

SolarWinds frames Avior as a way to extend the capabilities of existing teams rather than replace IT professionals. The agents can offer advice and also act on an administrator's instruction. That distinction is central to the product's design. An assistant or co-pilot can suggest a fix; an agent that can execute tasks can carry the fix through approved steps, record what happened and report back.

The value is most apparent in day-to-day operational work: clearing repetitive alerts, checking configurations, maintaining backups and following compliance policies. Those tasks are necessary but often pull engineers away from architecture, capacity planning, security improvement and other higher-value work. By handling some of that tactical load, Avior aims to give teams more time for projects that require human judgment and business context.

Rollout and Roadmap

Avior's network engineering agent is available now through a guided onboarding program. SolarWinds expects to introduce a database administration agent later in the fourth quarter, followed by IT administration, site reliability engineering and security analyst agents across late fourth quarter and early 2027.

The phased rollout suggests SolarWinds is starting with a domain where it has deep operational data and customer experience, then extending the agentic framework to adjacent IT functions. Each new agent will face different risk profiles and workflows. Database administration, IT administration, site reliability engineering and security analysis all involve distinct systems, permissions and compliance requirements. The common thread is the agentic layer: agents that can use operational data, execute workflows and operate within guardrails set by administrators.

For network teams evaluating Avior, the practical question is not whether agents can act, but where they should act. The control modes and safeguards give administrators a way to start with recommendations and approvals, then widen autonomy for repetitive or low-risk tasks. That graduated approach may prove as important as the automation itself, because trust in agentic AI will depend on predictable behavior, clear audit records and the ability to reverse course when something goes wrong.

Source:Network World News