Support — one scorecard, AI and humans as peersOwner: converging PRODUCT rows designed ONCE. The MAV-OWNED map lists Support as PRODUCT (Folo) ⇄ PRODUCT (Receipt) — the same job on two products, so this pack builds it as one engine-shaped common surface (the map’s own convergence rule). The shape is Folo’s live scorecard (walks §12) — the strongest support surface in the family. Ghost sessions in the inbox are read-first, banner-flagged, time-boxed and fully audited — the live discipline line, kept verbatim.
Scope (§17): one team, one scorecard, per-platform intake. Tickets arrive per platform (the intake chips), agents work ONE queue and rank on ONE scorecard — the scorecard is a cross-platform fact and says so; the ticket rows stay tagged with their intake platform. folo.link: composes all three intakes into this one queue; every mutation deep-links into the owning product’s admin — folo.link writes nothing.
Scope (§17): one team, one scorecard, per-platform intake. Tickets arrive per platform (the intake chips), agents work ONE queue and rank on ONE scorecard — the scorecard is a cross-platform fact and says so; the ticket rows stay tagged with their intake platform. folo.link: composes all three intakes into this one queue; every mutation deep-links into the owning product’s admin — folo.link writes nothing.
Team scorecard + ticket triage — AI agents and human agents rank in the same table.
Demo dataPartial · Folo live seed
Tickets handled · 28d
3,258
+4.2% vs prior window
DEMO DATA · window: 28d · source: ticket ledgerAI auto-resolve
49.7%
closed with no human touch
DEMO DATA · window: 28d · source: ticket ledgerFirst response · median
4.6m
all queues, AI + human blended
DEMO DATA · window: 28d · source: ticket ledgerResolve · median
1.2h
open → resolved
DEMO DATA · window: 28d · source: ticket ledgerCSAT
88
post-resolve survey score
DEMO DATA · window: 28d · source: surveySLA breach rate
1.1%
oldest open ticket: 19h
DEMO DATA · window: 28d · source: ticket ledgerEvery agent, one ranked tableAI agents and human agents are PEER ROWS — same columns, same ranking, badged AI / Human. The live Folo scorecard already does this (Orca and Atlas rank against Dana and Mara) and it is the pattern this family standardises on: an AI agent is a team member with a performance record, not a feature toggle.
Ranked by handled · 28d — the badge is the only difference between an AI row and a human row.
| Agent | Handled | FRT | Resolve | Auto / escalated | Hand-off acc. | CSAT | SLA breach | Backlog | |
|---|---|---|---|---|---|---|---|---|---|
| Orca L1 triage | AI | 1,412 | 0.4m | 0.6h | 71% / 29% | 93.1% | 86 | 0.7% | — |
| Atlas billing | AI | 494 | 0.6m | 0.9h | 58% / 42% | 90.2% | 84 | 1.0% | — |
| Dana R. | Human | 402 | 6.1m | 1.4h | — / — | — | 92 | 0.9% | 7 |
| Mara T. | Human | 361 | 7.4m | 1.7h | — / — | — | 90 | 1.2% | 5 |
| Jin H. | Human | 312 | 8.0m | 2.1h | — / — | — | 89 | 1.6% | 9 |
| Owen P. | Human | 177 | 9.2m | 2.4h | — / — | — | 87 | 1.9% | 4 |
| Kavi N. | Human | 100 | 10.8m | 2.9h | — / — | — | 88 | 1.4% | 6 |
≡
Auto/escalated and hand-off accuracy apply to AI rows only — a human row shows a dash, never a fake number.
AI vs human handling
50% AI auto
AI auto-resolved · 50%AI → human escalated · 26%Human-only · 24%
Hand-off accuracy
When AI escalates, does it hand the ticket to the right human queue? Misroutes cost a second FRT window.
Ghost sessions in this inboxA support agent may ghost into a user’s session to see what they see. The discipline is non-negotiable: read-first, banner-flagged, time-boxed, fully audited — every ghost entry from this inbox writes to the session log on ghost.html.
Read-first · banner-flagged · time-boxed · audited — session log →