Live · in pilot

Reactivation — stale → won, end-to-end.

Stale customer → first touch → reply → meeting booked → meeting completed → deal closed. Cortado owns the loop, you approve the sends, the ledger tracks the dollars.

Watch her work — live dashboard → Book a tasting

What the skill does, step by step

Watch

Stale-customer detection

Cortado queries your customer list daily. Anyone who hasn't been contacted in 60+ days (configurable per vertical) lands in the stale queue.

Notice

Insight generation

Cortado runs the pattern queries: how many stale, by segment, by value, by recency, by contact-confidence. Writes headline + detail + recommendation.

Draft

Channel-specific campaigns

Cortado drafts the outreach. Verified-email → personalized email per customer. Verified-phone → live-call queue with opener/question/booking-close. Inferred/unknown → LinkedIn intro-first.

Approve

Owner gate

Nothing sends without your ✓. Review drafts, approve the campaign (all at once or per-recipient). Cortado sends at the cadence you set.

Run

Loop execution

Cortado watches replies, classifies intent via Claude, books meetings via Cal.com, captures meeting outcomes. You only step in when a deal needs you.

Receipt

Ledger write

Every action with a dollar figure: first-touch sent, reply received (intent-adjusted value), closed-won (realized). Headline number updates in real time.

What enters the stale queue

🕰

Old customers

Past customers who haven't bought in N days (configurable per vertical)

📋

Quoted-but-not-closed

Anyone who got a quote but didn't accept within X days

📅

Booked-but-no-show

Meetings scheduled but missed; specific re-outreach pattern

🎯

Prospect hotlists

Pre-researched lead lists (the Cafecito-AI pilot is exactly this case)

💼

Lost deals

Closed-lost from N months ago, where the original objection might be stale

How dollar attribution works

Per closed-won customer

Cortado writes value_usd_realized = expected year-one revenue from this customer (setup fee + 12 × monthly OR your configured LTV per customer type). For Cafecito-AI's pilot, that's $34,000 per closed prospect.

Per pending stage (reply, meeting-booked)

Cortado writes value_usd_estimate = expected_value × close_rate_at_this_stage, tuned by intent classification.

Classified-interested reply: 25%. Info-request reply: 18%. Unclear reply: 10%. Unsubscribe / wrong-recipient: 0% (closed-lost).

The headline number

Only status='realized' AND owner_approved=1 rows count toward the $25k/90-day refund threshold. Estimates show separately so you see pipeline depth without inflating the contract number.

Live now

Watch Cortado work the Cafecito-AI pilot.

28 Miami prospects in the stale queue, real insights generated, three campaigns drafted (email + phone + LinkedIn), one closed-won loop seeded. No screenshots, no demo data.

Honesty doctrine: Reactivation is in pilot. It works against Cafecito-AI's own hotlist today. The "live" status applies only to this skill; phone, intake, quote, reviews, orders ship Q3 2026 with the same ledger + approval pattern.