Where the AI engineer hands off to a technician.
When the AI engineer escalates, the technician picks up an open ticket with investigation findings, attempted actions, and full endpoint context. The toolkit is the bench they continue the work on, governed by the same approvals, audit trail, and evidence layer.
Each tool has a job, a control boundary, and a receipt.
The toolkit is not a pile of buttons. Each surface exists for a specific support moment, and every action leaves evidence for the ticket.
Technician Chat
Ask the AI engineer to inspect the endpoint, explain findings, recommend the next step, request approval, run an approved action, and summarize what happened.
End-User Chat
Keep users informed, ask clarifying questions, offer approved L1 quick actions, and hand off to a technician without losing the ticket thread.
Remote Control
Join the user session when the issue needs eyes-on-screen support, guided steps, or visual confirmation that the fix worked. Remote control is available for Windows, Linux, and macOS endpoints.
File Transfer
Upload installers, retrieve logs, collect screenshots, or move vendor repair tools without switching to a separate support product.
Secure Terminal
Run diagnostics and escalations directly on the endpoint while command text, output, exit status, and timestamps are captured beside the ticket.
Native Actions
Use fixed, catalogued actions before generated commands: Flush DNS, clear browser cache, restart spooler, renew DHCP, restart services, and other known-safe operations.
Commands
Review command runs, native action executions, outputs, approvals, and failures so technicians can understand what already happened before rerunning anything.
Asset Info
See the endpoint identity, OS, ownership, policy context, organization, user, source connector, and health signals before taking action.
Processes
Inspect running processes, CPU and memory usage, suspicious spikes, stuck applications, and user-session activity before deciding what to restart or close.
Services
Review service state, startup type, recovery behavior, and recent failures, then start, stop, or restart services through governed actions.
Applications
Review installed applications, versions, publishers, install dates, and repair or removal candidates during software and browser support.
Updates
Check update history, pending reboots, failed update signals, and repair paths for Windows Update and supported operating system update flows.
Windows Event Viewer
For Windows endpoints, technicians can inspect relevant event logs to understand service crashes, update failures, driver errors, authentication issues, and application faults.
Timeline and Evidence
Every chat request, approval, terminal command, native action, file movement, verification check, and ticket update lands in one operational record.
Power tools with operational control.
Approval policies
Risky commands, destructive actions, production-impacting changes, and customer-specific rules pause for approval before execution.
Native before script
Known remediation uses catalogued native actions where possible, so teams get deterministic behavior before generated commands are considered.
Evidence by default
Every terminal run, quick action, investigation, workflow step, and escalation leaves output a technician can review later.
Manual support, AI support, and automated support use the same foundation.
Support tools are not separate from automation. They are the actions and evidence layer that resolution playbooks and custom workflows rely on.
AI chat can ask for input, suggest quick actions, or escalate to a technician.
Playbooks use native actions and verification checks to close known endpoint issues.
Workflows sequence investigations, decisions, approvals, actions, and closeout notes.
The ticket timeline keeps the user, technician, and approver on the same record.
Autonomy rate shows which issue classes are actually closed end to end.
Policies decide when support tools can run automatically and when approval is required.
Give the AI engineer a real support bench.
Start with the tools technicians already need, then let playbooks and workflows automate the parts that are safe to close.