/01
Support agent (Gorgias / Zendesk-aware)
Handles 60–80% of tickets autonomously — tracking, sizing, returns, exchanges — with policy awareness baked in. Edge cases route to a human with full context.
Four patterns we hear from e-commerce brands on every diagnostic call. None of them get fixed by hiring. All of them get fixed by writing the implicit rules down and letting AI carry the rote work.
Handles 60–80% of tickets autonomously — tracking, sizing, returns, exchanges — with policy awareness baked in. Edge cases route to a human with full context.
Policy-aware refund or store-credit decisions. Eligible items refunded automatically; ineligible items routed with a drafted explanation.
Triple Whale + Shopify + Meta + Google → branded daily report by 6am. The team stops assembling; the team starts deciding.
Sell-through rate vs runway. Slack alerts when a SKU will stock out before the next supplier delivery. POs drafted on detection.
PO sent → tracking pulled → ETA updates pushed to the buying team. Late shipments surface days earlier.
Sized items get review asks at 14 days. Consumables at the half-empty point. Cadence aware of what the customer actually bought.
Pause, swap, skip — offered before cancel. Churn drops; voluntary cancels become save opportunities.
| # | Manual today | With Scooper |
|---|---|---|
| /01 | Customer support volume scales with revenue | Support agent (Gorgias / Zendesk-aware) |
| /02 | Returns and refunds drain margin and time | Returns automation |
| /03 | Reporting is a manual mess | Daily ad performance + finance report |
| /04 | Inventory and supplier coordination is reactive | Inventory anomaly detection |
We do not migrate you off your software. We build a layer on top.
“Support response time under 1 minute. CX team handles 3× volume. Inventory misses get caught the same day they happen, not the day before stock-out.”
15-min discovery → two-week paid diagnostic. Operator interviews, workflow shadowing, and three nasty real examples worked end-to-end. By day ten you have a scope, a price, and a signed-off KPI.
Senior operators in your repo and your Slack. Code, prompts, and evals shipped in 3–6 weeks. KPI signed off in writing before we start. Miss the number, the next sprint is on us.
Two rounds of post-launch tuning. Then your team owns it. We hand over the code, the eval harness, the runbook, and the SOPs. No retainer required. No vendor lock-in.
Anonymized · numbers from the books
A $14M DTC consumables brand had a 6-person CX team and a returns problem that ate 40% of margin. We shipped the support agent in four weeks and returns automation in six. Auto-resolved tickets went from 0 to 71%, return processing time dropped 90%, and the next $4M in revenue came without a CX hire.
We will tell you in 15 minutes whether AI fits your bottleneck. If it does, we will scope a two-week paid diagnostic. If it does not, we will tell you that too.