Put your best network engineer on every shift
Selector Foundry is the Agentic NetOps platform on Selector’s full-stack observability and AIOps foundation. Its agents correlate the evidence, find the root cause, act inside your guardrails, and prove recovery, all on a single correlated model of your whole path.
Six specialists behind one conversation
Selector Foundry puts a team of specialized agents on the context the Selector platform already builds. Each one owns a single field of expertise and a defined set of tools. You ask Rosetta in plain language, and it works out which specialists answer.
What's going on with the Denver branch?
Rosetta
Reads the question, calls the specialists it needs in parallel, and returns one answer with the evidence attached.
Root Cause
Loom
Weaves the signals into the root cause that started it, with a trail you can follow.
ITSM
Ticketmaster
Opens and updates the ticket in ServiceNow or Jira, without any manual work.
Notifications
Herald
Sends the right update to the right channel, the moment it matters.
Dashboard
Artist
Builds the exact view the question needs, so no one hunts across dashboards.
Agent Builder
Sculptor
Builds new specialists for your custom workflows, inside your guardrails.
Rosetta · interface & orchestration
Run the play
Ask in plain language. Rosetta is your single interface to the platform, and the orchestrator that calls the right agents and returns one answer.
- Ask: put any question in plain language, no query syntax or dashboards needed.
- Orchestrate: reach every specialist agent through one conversational interface.
- Continue: keep the thread going across turns; Rosetta holds the context.
- Inspect: open the reasoning, the agents it called, and the evidence behind every claim.
Loom · correlation + RCA
Connect the dots
Loom weaves a flood of alerts across every domain into a few correlated incidents, telling source from symptom and naming the root cause.
- Correlate: collapse a storm of alerts across tools and domains into a few real incidents.
- Isolate: tell the causal fault from its downstream symptoms, so you chase one, not many.
- Reconstruct: build the timeline, blast radius, and evidence behind every incident.
- Anticipate: see the next risk on the path before it becomes its own incident.
TicketMaster · ticketing + ITSM
Manage the tickets
TicketMaster files the incident in your own ITSM with the root cause and evidence attached, routes it to the right owner, and keeps it current as things change.
- Create: open an incident in ServiceNow or your ITSM of choice, root cause included.
- Enrich: assign it to the right owner and ping them in Slack, Teams, or anywhere else you work.
- Route: attach the evidence, context, and impact so no one starts cold.
- Update: keep the ticket current as the investigation changes, both directions.
Herald · notifications
Keep everyone in the loop
Herald turns a finding into the right update for every audience and delivers it to the channels your team already uses, on demand or on a schedule.
- Tailor: shape one finding into an executive brief, an ops note, or a customer status.
- Publish: post it to Slack, Teams, email, or wherever the audience already is.
- Schedule: send recurring updates, like a Monday briefing, on a set cadence.
- Document: produce the post-incident review with a verified narrative and timeline.
Artist · dashboards on demand
Build the view
Artist turns a plain-language request into a live dashboard, choosing the right widgets and laying them out the way you asked, in seconds.
- Describe: ask for a dashboard in plain language and get it built in seconds.
- Visualize: turn each metrics into the right widget, heatmap, honeycomb, table, or graph.
- Arrange: place the widgets where you want them, side by side or full width.
- Refine: tweak any widget on the fly, including the window, the chart type, or the layout.
Sculptor · agent builder
Build your own agents.
Sculptor turns your playbooks into agents. Describe what you want in plain language, and it builds, triggers, and runs a governed agent in the background.
- Describe: spell out the agent you want in plain language, and Sculptor picks the tools, sets it up, and builds it.
- Trigger: Run it on an event, a schedule, or an API call, so it fires exactly when the work needs to happen.
- Track: watch its runs, tasks, and logs as it works in the background, with full visibility into everything it does.
- Govern: version it in your own GIT repo, under your RBAC, scoped to only the tools you allow.
Turn a 250-person war room into a six-agent fix.
Foundry compresses the incident lifecycle, from fragmented signals to an evidence-backed resolution, so your best engineers spend their time on the network instead of the bridge call.
Correlation and deterministic root cause
Root cause is named before the bridge call assembles. MTTR drops from four hours to thirty minutes, an 8x improvement.
One correlated model of the full path
One team resolves the incident instead of every domain proving it is innocent. War rooms shrink from 250 engineers to 5.
Agents run all five moves
Engineers stop hand-carrying triage between tools. Roughly 1,000 engineer-hours are reclaimed per major incident.
Event-led operation
The investigation starts before anyone is paged, so incidents are caught rather than chased.
Governed action with approval gates
Remediation your team can trust. A human approves every change, inside the limits you set.
Vendor-neutral collection, 300+ sources
No rip and replace. Add a domain or a vendor without re-architecting anything.
Correlation and deterministic root cause
Root cause is named before the bridge call assembles. MTTR drops from four hours to thirty minutes, an 8x improvement.
One correlated model of the full path
One team resolves the incident instead of every domain proving it is innocent. War rooms shrink from 250 engineers to 5.
Agents run all five moves
Engineers stop hand-carrying triage between tools. Roughly 1,000 engineer-hours are reclaimed per major incident.
Event-led operation
The investigation starts before anyone is paged, so incidents are caught rather than chased.
Governed action with approval gates
Remediation your team can trust. A human approves every change, inside the limits you set.
Vendor-neutral collection, 300+ sources
No rip and replace. Add a domain or a vendor without re-architecting anything.
Selector Recognized in the 2026 Gartner® Market Guide for Agentic NetOps Software
Read Gartners latest resport on Agentic NetOps Software to understand how the market is evolving, which capabilities matter most, and why Selector is recognized as a representative vendor in this emerging market.
Governed by design
Foundry agents run where your telemetry already lives. You author them as software, version them in your own Git, and run them under the controls you already enforce, so nothing is copied to an outside service.
Runs on your data
Agents execute where your telemetry lives. Nothing leaves.
Authored as software
Written by your team, committed to your Git, reviewed via pull requests.
Validated before promotion
Replayed against past incidents, with one-step rollback.
Under your controls
Run on your LLM of choice, under your RBAC and identity policies.
News & insights
Every network engineer carries an instinct that outsiders mistake for caution: a change in one place can travel. Reroute a path, push a policy, drop an interface, and the effect
