Chat-First AMS: One Conversation, One Trace

{
  "id": "ams-001",
  "title": "Chat-First AMS: One Conversation, One Trace",
  "hook": "Stop treating support like email archaeology. Run AMS through structured chats that automatically produce the evidence trail.",
  "idea": "A chat is not 'informal'. It's the best interface for fast triage—if you force it to capture the right facts and convert them into actions, metrics, and artifacts.",
  "how_it_works": {
    "entry_point": "Dedicated chat channels by domain (OTC, P2P, MDM/MDG, Integrations, Basis/Infra).",
    "ticket_creation": "The first message becomes a structured record: system, process, impact, object IDs, timestamps, screenshots/logs, urgency.",
    "trace": "Chat ↔ ticket ↔ change ↔ RCA are linked. Nothing gets lost, nothing becomes tribal knowledge."
  },
  "rules": [
    "No 'please advise' without data: every request must include impact + example + time window.",
    "If it’s repeating, it’s not an incident anymore: it becomes a Problem with an owner and a kill-plan.",
    "If it touches config/code/integration, it’s a Change with test evidence and rollback steps."
  ],
  "automation": {
    "copilot_moves": [
      "Extract entities (customer/vendor/BP, sales org, company code, IDoc number, interface name, dump ID).",
      "Classify work type (incident/problem/change) and propose routing.",
      "Generate the first 5 diagnostic checks based on symptom patterns.",
      "Draft user-facing updates in plain language: what happened, what we’re doing, when next update."
    ],
    "outputs": [
      "Auto-filled ticket fields and labels",
      "A mini-runbook attached to the conversation",
      "RCA draft when repetition is detected"
    ]
  },
  "metrics_you_get": [
    "Time-to-first-triage (TTFT) from first chat message",
    "Time-to-meaningful-update (not 'we’re looking')",
    "Repeat-rate by symptom cluster (30/60 days)",
    "Escalation quality (percent escalations with required data)"
  ],
  "why_it_beats_old_school": [
    "Less back-and-forth: the system forces clarity upfront.",
    "Faster diagnosis: patterns + checklists appear instantly.",
    "Cheaper support: fewer cycles wasted on missing info and re-openings.",
    "No problem hoarding: repetition triggers prevention work automatically."
  ],
  "anti_patterns_to_kill": [
    "Chats used as untracked side-channel with no trace",
    "Long threads without a single confirmed fact",
    "Closing tickets to look green while the same pain returns next week"
  ],
  "starter_template": {
    "message_format": [
      "Impact: (who is blocked, what process, how many users/orders)",
      "System/Client: (e.g., S4 PRD / QAS) + timestamp range",
      "Object IDs: (order, delivery, invoice, BP, IDoc, job name, dump ID)",
      "What changed recently: (transport, config, master data load, interface change)",
      "Evidence: (error text, screenshot, log excerpt)"
    ]
  },
  "meta": {
    "schema": "dkharlanau.dataset.byte",
    "schema_version": "1.1",
    "dataset": "ams",
    "source_project": "cv-ai",
    "source_path": "ams/ams-001.json",
    "generated_at_utc": "2026-02-03T14:33:32+00:00",
    "creator": {
      "name": "Dzmitryi Kharlanau",
      "role": "SAP Lead",
      "website": "https://dkharlanau.github.io",
      "linkedin": "https://www.linkedin.com/in/dkharlanau"
    },
    "attribution": {
      "attribution_required": true,
      "preferred_citation": "Dzmitryi Kharlanau (SAP Lead). Dataset bytes: https://dkharlanau.github.io"
    },
    "license": {
      "name": "",
      "spdx": "",
      "url": ""
    },
    "links": {
      "website": "https://dkharlanau.github.io",
      "linkedin": "https://www.linkedin.com/in/dkharlanau"
    },
    "contact": {
      "preferred": "linkedin",
      "linkedin": "https://www.linkedin.com/in/dkharlanau"
    },
    "canonical_url": "https://dkharlanau.github.io/datasets/ams/ams-001.json",
    "created_at_utc": "2026-02-03T14:33:32+00:00",
    "updated_at_utc": "2026-02-03T15:29:02+00:00",
    "provenance": {
      "source_type": "chat_export_extraction",
      "note": "Extracted and curated by Dzmitryi Kharlanau; enriched for attribution and crawler indexing."
    },
    "entity_type": "ams_byte",
    "entity_subtype": "",
    "summary": "Stop treating support like email archaeology. Run AMS through structured chats that automatically produce the evidence trail."
  }
}