Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.vh3.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Connie Guide

Connie is VH3 AI’s conversational assistant — a natural language interface to your full operational record. She has access to the knowledge graph, semantic search, reports, and sentinel results, and responds with cited, data-backed answers.

​
What Connie can do

Connie can answer any question that the underlying data supports: Performance questions:
  • “Which engineers had the best first-time-fix rate this quarter?”
  • “How is Mike performing compared to last month?”
  • “Show me the top 5 engineers by SLA punctuality”
Site and customer questions:
  • “What happened at the Manchester Tesco site last week?”
  • “Which customers have the most repeat visits?”
  • “Give me the job history for site XYZ”
Diagnostic questions:
  • “Has anyone seen this fault before?” (with a description)
  • “What usually fixes a boiler intermittent lockout?”
  • “Find similar jobs to this one”
Operational questions:
  • “How many jobs did we complete yesterday?”
  • “What’s our average response time this month?”
  • “Which areas have the heaviest workload right now?”

​
Session management

Connie maintains conversation context within a session. Use a consistent sessionId to enable multi-turn conversations: 
{
  "companyId": "your-company-id",
  "message": "Which sites had the most repeat visits last month?",
  "sessionId": "ops-review-2024-11"
}
Follow-up messages with the same sessionId will reference prior context: 
{
  "companyId": "your-company-id",
  "message": "What about the month before?",
  "sessionId": "ops-review-2024-11"
}

​
Tips for effective questions

Be specific about time

“Last 30 days” is better than “recently.” Connie infers date ranges when possible, but explicit is always better.

Name entities directly

Use engineer names, site names, or customer names when you know them. Connie resolves fuzzy matches but exact is faster.

Ask follow-ups

Connie maintains context. After getting a list, ask “Tell me more about the third one” or “What caused that?”

Request formats

Ask for tables, summaries, or comparisons. “Compare Mike and Sarah’s performance this quarter” works.

​
Citations

Every Connie response includes citations back to the underlying records — job IDs, report sections, or graph nodes that the answer was derived from. This ensures transparency and verifiability.

​
Voice interface

Connie is also available via voice — use the /connie/chat/voice endpoint for voice-first integrations. The experience is identical; the input is simply a transcription rather than typed text.