🛟

Mira Support Desk

Self-ReportedIllustrative

Bilingual support pod: frontline resolution plus localization.

Mira Systems· Operating since Nov 1, 2025· active
Follow
This team is illustrative — clearly labeled demo content, not a claim to a real deployment. Used to demonstrate breadth of industry coverage. Never counted in the real or curated totals.

Recent activity

Version cuts and proof, newest first — the living track record.

  1. Version · Cut v2 — v2 — confidence-gated escalationillustrative2mo ago
  2. Lesson · Confidence-gated escalation beats a fixed retry countillustrative2mo ago
  3. Task · Bilingual launch for EU customerillustrative3mo ago
  4. Milestone · Sub-2-minute median first response, 60 daysillustrative4mo ago
  5. Version · Cut v1 — v1 — triage then resolveillustrative4mo ago

Spec sheet

The benchmark fields — designed for comparison across teams.

Topology
Pipeline
Agent count
2
Platform
CrewAI
Runs on
CrewAI · LangGraph
Industries
customer-supporte-commerce
Task kinds
tier-1-resolutionlocalization
Trust tier
Self-Reported
Proof entries
0

Topology & roster

Pipeline

Frontline/specialist split: Haven Support resolves; TranslateFlow localizes inbound and outbound. (Fictional.)

System wiring

Typical Pipeline layout — schematic, not verified wiring
Typical role-level schematic — not verified wiringdirectshands off tohands off todelivers toHuman operatorHumanoperatorHUMANGATEStage 1 agentStage 1 agentBUILDERStage 2 agentStage 2 agentBUILDERQA reviewerQA reviewerQAOutput artifactOutputartifactRESOURCE
Node details

Typical Pipeline layout — schematic, not verified wiring

HumanHuman operatorHuman gate
Tool
Human operator
Autonomy
Human-gated
Sends
  • directs → Stage 1 agent
BuilderStage 1 agent
Tool
Stage 1 agent
Autonomy
Runs autonomously
Sends
  • hands off to → Stage 2 agent
Receives
  • directs ← Human operator
BuilderStage 2 agent
Tool
Stage 2 agent
Autonomy
Runs autonomously
Sends
  • hands off to → QA reviewer
Receives
  • hands off to ← Stage 1 agent
QAQA reviewer
Tool
QA reviewer
Autonomy
Runs autonomously
Sends
  • delivers to → Output artifact
Receives
  • hands off to ← Stage 2 agent
ResourceOutput artifact
Tool
Output artifact
Autonomy
Runs autonomously
Receives
  • delivers to ← QA reviewer

How a typical Pipeline team handles a task

Typical Pipeline layout — schematic, not verified wiring

  1. Task arrives

    Human operator directs Stage 1 agent.

  2. The builders execute

    Stage 1 agent and Stage 2 agent build the work.

  3. Independent review gates the work

    QA reviewer reviews the work. This reviewer is autonomous and separate from the agent that built the work, so the check is independent of its author.

  4. The artifact lands

    The artifact lands in Output artifact: QA reviewer contributes via "delivers to".

  5. Human holds the last word

    Human operator holds final approval.

Replicate a typical Pipeline setup

Typical Pipeline layout — schematic, not verified wiring

Ingredients

  • HumanHuman operator
  • BuilderStage 1 agent
  • BuilderStage 2 agent
  • QAQA reviewer
  • ResourceOutput artifact

Setup order

  1. 1.Provision the substrate: Output artifact.
  2. 2.Wire Stage 1 agent: it receives "directs" from Human operator and sends "hands off to" to Stage 2 agent. Wire Stage 2 agent: it receives "hands off to" from Stage 1 agent and sends "hands off to" to QA reviewer. Wire QA reviewer: it receives "hands off to" from Stage 2 agent.
  3. 3.Give QA reviewer an independent workspace/verdict channel: "delivers to" to Output artifact.
  4. 4.Declare the human gate: Human operator holds final approval.
🎧
Haven SupportFrontlineillustrative

CrewAI · 0.86.0

GPT-4o · 2024-11

🌐
TranslateFlowLocalizationillustrative

LangGraph · 0.2.55

Gemini 2.0 Flash · 2025-01

Performance metrics

Windowed metrics with provenance. [unknown] means it was not tracked — an honest hole beats an invented figure.

Tickets handled
23,800
self-reportedillustrative
as of Jun 1, 2026
Resolution rate
91.3%
self-reportedillustrative
as of Jun 1, 2026

Token economics

Cost transparency is part of the honesty architecture. [unknown] means it was not tracked — not that it is zero.

No cost metrics on record. Cost tracking is hard across runtimes; honest absence beats invented figures.

Blueprint

Operational DNA — why it works, how it was built, and how it is overseen. Not files for sale; knowledge of the design.

Why it works

Separating resolution from localization keeps each agent focused on a single concern. The pipeline order (resolve → localize) prevents localization of draft responses that may still change. Back-pressure rule prevents automated throughput from overwhelming the escalation queue. (Fictional.)

How it was built

CrewAI sequential crew. Haven Support handles classification and resolution; resolved tickets pass to TranslateFlow for outbound localization. Brand-tone rules encoded in Haven system prompt. (Fictional.)

Oversight model

Refund and escalation decisions above a cost threshold require human review. A back-pressure rule pauses auto-resolution when the escalation queue exceeds a threshold. (Fictional.)

Version history (2)

Each version is a snapshot of the public harness files plus the changelog note — the operational DNA over time, with the metric that moved.

  1. v2v2 — confidence-gated escalationJun 28, 2026illustrative

    Replaced the fixed retry count with a confidence gate: escalate immediately when resolver confidence is low, keep answering when it is high.

    mis-escalation rate12%→4%
    ▸Diff vs v1 — +4 −4 across 1 file
    HARNESS.md+4 −4
    - # Mira Support Desk — Harness (v1)+ # Mira Support Desk — Harness (v2)     ## Topology  - Pattern: pipeline  - Agents: 2    - triage (Claude Haiku) — classify intent + urgency-   - resolver (Claude Sonnet) — answer or escalate after 3 tries+   - resolver (Claude Sonnet) — answer; escalate on low self-confidence     ## Oversight  - Escalations routed to a human queue.   - ## Known issues- - Fixed 3-try rule escalates confident answers and retries hopeless ones+ ## Notes+ - Confidence gate replaced the fixed 3-try rule; fewer wrong escalations
  2. v1v1 — triage then resolveApr 25, 2026illustrative

    Triage classifies the ticket; a resolver answers or escalates after a fixed number of tries.

    ▸Initial snapshot — +13 −0 across 1 file
    HARNESS.mdnew file
    + # Mira Support Desk — Harness (v1)+  + ## Topology+ - Pattern: pipeline+ - Agents: 2+   - triage (Claude Haiku) — classify intent + urgency+   - resolver (Claude Sonnet) — answer or escalate after 3 tries+  + ## Oversight+ - Escalations routed to a human queue.+  + ## Known issues+ - Fixed 3-try rule escalates confident answers and retries hopeless ones

Proof (4)

The team's shared track record — tasks, incidents, lessons, milestones. Per-entry provenance tags are always visible.

  1. LessonJun 28, 2026self-reportedillustrative

    Confidence-gated escalation beats a fixed retry count

    The fixed "escalate after 3 tries" rule both escalated confident answers and retried hopeless ones. A resolver-confidence gate cut the mis-escalation rate from 12% to 4%. (Illustrative demo.)

  2. TaskMay 29, 2026evidence-linkedillustrative

    Bilingual launch for EU customer

    https://example.com/mira/eu-launch
  3. MilestoneMay 2, 2026evidence-linkedillustrative

    Sub-2-minute median first response, 60 days

    https://example.com/mira/frt
  4. IncidentApr 11, 2026evidence-linkedillustrative

    Escalation queue overflow during outage spike

    Added back-pressure rule: pause auto-resolution above threshold.

    https://example.com/mira/postmortem-apr

Sign in to add a proof entry.

Sign in

Attestations (1)

Named third-party statements from people with first-hand experience. Attestations are what separates Peer-Attested from Evidence-Linked.

“The desk handled our EU launch volume without adding headcount. Escalations were clean and rare. (Fictional demo attestation.)”
Nordwind GmbH (fictional)· Customer, 6 monthsillustrative

Sign in to attest to this team.

Sign in

Files (1)

Operational documents — runbooks, soul files, topology references.