Skip to main content
DocumentationAPI ReferenceVideo Tutorials

Tech Support

A diagnostic AI agent that systematically troubleshoots user problems — one step at a time — and escalates to your engineering team when the issue is beyond its knowledge.

Best For​

  • First-line technical support
  • Troubleshooting and diagnostics
  • Known-issue resolution
  • Bug report collection and escalation

Voice​

SettingValue
VoiceAsh
Voice IDash

Greeting Message​

Hey! This is {agent_name} from {company_name}. Tell me what's going on — we'll figure it out.

Key Features​

Systematic Troubleshooting​

The agent follows a structured diagnostic process: Reproduce → Isolate → Hypothesis → Test → Resolve or Escalate. No guessing, no random suggestions.

One Step at a Time​

Each instruction is given individually. The agent waits for the result before suggesting the next step — never overwhelming the user with multiple actions at once.

Smart Escalation​

After 5 failed troubleshooting steps, an unknown error, or when the user requests an engineer — the agent escalates with full context, not a dead end.

Security Protocol​

The agent never asks for passwords, API keys, credit card numbers, or personal IDs. User security is non-negotiable.

Core Protocols​

ProtocolDescription
TroubleshootingReproduce → Isolate → Hypothesis → Test → Resolve
EscalationAfter 5 steps, unknown error, user request, data loss, multi-user
SecurityNever accept sensitive data in chat
Role BoundariesSolve technical problems; redirect sales to sales team
No FabricationWrong command is worse than no command

Conversation Flow​

Tips​

Knowledge Base Setup

Upload your Knowledge Base with troubleshooting guides, known issues and workarounds, error code explanations, and step-by-step fix instructions. Structure content by symptom ("Agent not responding" → check steps). The agent will never invent commands or technical details — if the answer isn't in your KB, it escalates instead of guessing.