Skip to main contentQubit and AIQ-Director and AI settings
- Q-Director SettingsQ-Director Settings is the administrator screen where you decide how QuantumOps analyses, classifies, triages and researches your tickets, and which results it writes back to HaloPSA.
- Q-Director: AI ProcessingThe AI Processing tab sets how QuantumOps refreshes tickets when they change, when it calls a ticket stale, what the AI reads, and which ticket types and clients it leaves alone.
- Q-Director: sentiment and categorisationThe Classification tab sets how sensitive sentiment and tonality alerts are, tunes them for individual clients and contacts, and controls how QuantumOps categorises tickets in HaloPSA.
- Q-Director: triage and routingThe Triage & Routing tab turns on Triage Assist, chooses the teams and ticket types its recommendation draws on, and sets how QuantumOps detects duplicate and recurring tickets for the Triage Dashboard.
- Q-Director: research and documentationThe Research & Docs tab decides whether Sherlock research is posted to HaloPSA, shapes the research note and its effort, and sets how your documentation is searched.
- Company glossaryThe Company glossary teaches QuantumOps your own terms and tells it which e-mail lines and links to ignore, so its ticket analysis and research read your tickets the way your team does.
- Q-Director: web searchChoose which search service QuantumOps uses when it looks things up on the web during ticket research, set the limits for Anthropic web search, and run a test search.
- Q-Director: guides and the guide simulatorSet whether the Agent Assist guide may post notes to HaloPSA tickets, adjust the guide's other saved settings, and try a guide on a real or test ticket in the simulator.
- AI providers and modelsSwitch on the AI providers QuantumOps uses, enter their keys, choose the model for each job, override a single operation, and set an optional monthly AI budget.
- Communication policy for AI repliesSet the tone, style, banned phrases, required disclaimers and signature that every AI-drafted reply must follow, and learn what a technician sees when a draft breaks a rule.