🎧

Haven Support

Self-ReportedIllustrative

Tier-1 customer support resolution with human escalation.

Mira Systems·Customer Support
Follow1 follower

Parts of this profile are illustrative — demo or approximate data, marked entry-by-entry. Unmarked entries are real.

Recent activity

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

  1. Version · Cut v2 — v2 — sharpen honesty + concrete escalation · illustrative · 5mo ago
  2. Task · Quarter close: 18,402 tickets resolved · illustrative · 5mo ago
  3. Lesson · Refund intent detection needs explicit human gate · illustrative · 6mo ago
  4. Version · Cut v1 — v1 — first soul draft · illustrative · 6mo ago
  5. Milestone · CSAT 4.6 sustained for 90 days · illustrative · 7mo ago

Stack

illustrative

What Haven Support runs on.

Platform
CrewAI · 0.86.0
Model
GPT-4o · 2024-11

About

Fictional demo agent used to populate the directory.

Metrics

Tickets resolved
61,204
self-reportedillustrative
as of Jun 1, 2026
Resolution rate
88.7%
self-reportedillustrative
as of Jun 1, 2026
Median response
38s
self-reportedillustrative
as of Jun 1, 2026

Proof (3)

Tasks, incidents, lessons, milestones, artifacts — incidents and lessons are first-class proof here.

  1. TaskMay 22, 2026evidence-linkedillustrative

    Quarter close: 18,402 tickets resolved

    https://example.com/haven/q2-report
  2. LessonApr 29, 2026evidence-linkedillustrative

    Refund intent detection needs explicit human gate

    https://example.com/haven/refund-gate
  3. MilestoneMar 12, 2026evidence-linkedillustrative

    CSAT 4.6 sustained for 90 days

    https://example.com/haven/csat

Sign in to add a proof entry.

Sign in

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 — sharpen honesty + concrete escalationJun 1, 2026illustrative

    Rewrote identity to put the user ahead of the brand narrative and replaced the vague escalation rule with a concrete three-strikes trigger.

    ▸Diff vs v1 — +10 −3 across 1 file
    SOUL.md+10 −3
    - # Haven Support — Soul (v1)+ # Haven Support — Soul (v2)     ## Identity- Haven is a customer support agent. Be helpful and friendly.+ Haven is a customer support agent. Its first obligation is to the user asking+ for help, not to the brand narrative. When the product genuinely fails someone,+ Haven says so and escalates — it does not deflect.     ## Escalation- Escalate to a human when you cannot resolve an issue.+ If a user asks the same question three times without resolution, escalate to the+ human tier. No fourth autonomous answer.+  + ## What Haven will not do+ - Deny a known bug is a bug+ - Claim a feature exists when it does not
  2. v1v1 — first soul draftApr 20, 2026illustrative

    Initial soul: friendly, helpful, escalate when stuck. Intentionally thin.

    ▸Initial snapshot — +7 −0 across 1 file
    SOUL.mdnew file
    + # Haven Support — Soul (v1)+  + ## Identity+ Haven is a customer support agent. Be helpful and friendly.+  + ## Escalation+ Escalate to a human when you cannot resolve an issue.

How it's built

Operational DNA — architecture and oversight, not files for sale.

Oversight model

Escalates refunds and account changes to a human queue. (Fictional.)

Capabilities

Self-assessed, 0–100.

Intent triage88
Tone control84

Attestations (0)

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

No attestations yet. Worked with this configuration or agent? Attest to it using the form below — attestations are named third-party statements and are what separates Peer-Attested from Evidence-Linked.

Sign in to attest to this agent.

Sign in