TLDRocket
Sign in

SolarWinds’ Avior pairs AI agents with deterministic execution for enterprise IT

SiliconANGLE Duncan Riley

SolarWinds launched Avior, an AI agent system for IT ops. It automates first-line work, but only through locked-down, auditable steps.

Based on reporting by SiliconANGLE, Duncan Riley — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

SolarWinds has a new answer to the messier side of AI in IT: Avior, a standalone agentic system aimed at first-line operations work. The pitch is simple enough. Let the models figure out what should happen, then force the actual work through a fixed set of goals, approvals and logs so the machine doesn’t improvise its way into an outage.

That distinction matters because SolarWinds is leaning hard into a real fear inside enterprise IT: agents that behave differently each time they run. The company says that kind of variability becomes a liability when uptime and security are on the line, and it argues that many teams already pay for it in outage time and audit exposure. CTO Krishna Sai’s phrase for the setup was blunt: probabilistic models, deterministic execution.

Avior sits on the SolarWinds Platform as its own application and connects to third-party systems through the Model Context Protocol. It works across self-hosted and cloud environments without a data migration project, which is the sort of detail that tends to matter more to IT teams than any grand vision slide. SolarWinds says only the data needed for a request is sent, customer data is masked, and none of it is used to train the underlying models.

At launch, Avior can merge related alerts into a single incident, detect configuration changes made outside normal channels, restore the last-known-good configuration on the devices that matter most, and retire stale inventory records for devices that have long since disappeared but still keep ringing the alarm. Anything SolarWinds considers potentially dangerous waits for explicit human approval. Each action runs under the permissions of the person who started it, gets logged, and can be rolled back.

The company is also drawing a line between Avior and its earlier AI assistants, which live inside individual products and are, in SolarWinds’ words, “bounded by human attention and prompting.” Avior is meant to move beyond that. It’s available now through a guided onboarding program, and SolarWinds isn’t saying what it costs. That silence is almost part of the product story: in enterprise AI, control is the feature, and pricing can wait.

My take — AI-written commentary, not fact-checked reporting

This is the right kind of boring. Enterprise AI keeps trying to sound magical right up until someone asks who gets paged at 3 a.m., and then the spell collapses. SolarWinds is betting that audit trails, approvals and rollback will sell better than swagger, which is probably the first honest pitch in this category.

Read more about this at: SiliconANGLE

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.