PlatformAI investigations
From alert to likely cause in minutes.
When something breaks, Televane collects the related alerts, anomalies, error logs, traces and recent changes, ranks the likely causes along your service dependencies, and proposes next steps for you to approve.
What is an AI-assisted investigation?
When an incident starts, most time goes into gathering evidence: which alerts fired, what changed, which logs show errors, which dependency slowed down. An AI-assisted investigation does that gathering for you and ranks the likely causes.
Televane's investigations are read-only and evidence-based. They show findings with a confidence that never claims certainty, and any action they propose waits for a person to approve it.
Investigations (preview)
Start one from an incident, a service, a host or a plain question.
- Read-only analysis of alerts, anomalies, topology, logs, traces, Kubernetes, hosts, databases, checks and recent configuration changes
- Findings and ranked root-cause hypotheses, with a confidence that never claims certainty
- Optional automatic investigation for every new incident
Investigation — checkout latency
You stay in control
Suggestions, not surprises.
- Proposed actions — add a note, acknowledge, silence for an hour, rerun a check — each needs a person's approval
- Assistant chat in the portal for questions about your systems
- Uses your own OpenAI or Anthropic key through the AI gateway, if you turn it on; evidence is redacted and prompts are not stored
Get started in three steps
- Create a free account — Sign up at portal.televane.com. The Free plan needs no card, and new organizations can try Pro free for 14 days.
- Send data and connect alerts — Investigations work from the alerts, logs, traces and changes Televane already has.
- Start an investigation — Open one from an incident, a service or a host, or ask a question in plain words.
Works together with
AI investigations: common questions
Does the AI make changes on its own?
No. Investigations are read-only. Proposed actions, such as acknowledging or silencing an alert, each need a person's approval.
Which AI models does it use?
Your own OpenAI or Anthropic key through the Televane AI gateway, if you turn it on. Evidence is redacted and prompts are not stored.
Try it on your own systems
Create a free account, install the agent with one command and see your data in minutes. 14-day Pro trial, no card needed.
Start free