Built for $1M-$50M DTC brands. Scales with your ticket volume.
A custom-built AI Support fellow that handles routine tickets in your brand voice. A Customer Success fellow that watches retention and surfaces winback moments. A Copywriter fellow for product descriptions and email. All in your Slack.
Ecommerce has a specific pain shape. Ticket volume scales with sales \u2014 a viral moment can 5x support volume in a week. Customer expectations on tone are unforgiving \u2014 your brand voice in a return-policy reply matters as much as your homepage hero. Retention math is brutal \u2014 a 5-point lift in repeat-purchase rate is the difference between profitable and unprofitable.
Generic chat AI does not solve this. Macros and canned responses solve part of it but flatten brand voice. A FellowHire Support fellow is custom-built for your brand \u2014 trained on your voice, your return policy, your product catalog, your common ticket patterns. Lives in your helpdesk and your Slack. Drafts replies that sound like a senior CX person at your brand wrote them.
Honest scope: the fellow drafts; a CX lead reviews and sends (default). For narrow ticket categories you trust, we can configure auto-send. The brand voice stays yours.
Most ecommerce brands start with a Support fellow. Many add a CS fellow once retention becomes the focus.
Triages incoming tickets, drafts replies in your brand voice for common issues (returns, shipping, product questions), escalates the rest with full context.
Best fit if your ticket volume is the bottleneck.
Tracks repeat-purchase health, surfaces winback moments, drafts retention emails for VIP cohorts.
Best fit if retention is your focus.
Drafts product descriptions, email campaigns, ad copy. Trained on your brand voice.
Best fit if you need content that sounds like your brand.
Weekly metrics post, inventory anomaly alerts, vendor reconciliation.
Best fit if your ops team is past 10 people.
Patch reads incoming tickets across helpdesk, Instagram DMs, and email. Drafts replies for common categories (return status, order tracking, sizing questions).
Patch is trained on your tone \u2014 friendly, formal, snarky, whatever your brand uses. Drafts match.
Drafts the response, attaches the next-step instructions. CX reviews and sends, or auto-sends for narrow categories you trust.
Patch pings when high-value orders, suspicious orders, or repeat-customer issues land.
Sage builds repeat-purchase scores per cohort, flags when retention drops.
Sage surfaces customers who have not repurchased in 90/180 days, drafts winback emails for the marketing team.
River drafts product descriptions for new SKUs, returns 3 versions per length tier.
Rowan posts revenue, AOV, repeat rate, ticket volume, NPS to the founder channel every Friday.
We don't have an ecommerce case study to point to yet. We're being upfront about that. The Support and CS fellows are mature. Pilot customers in SaaS and MSP verticals are running similar workflows successfully today. Brand-voice fidelity is the biggest concern ecommerce buyers raise; we tune that during scoping with samples of your existing replies.
If your brand is open to being the anchor case study for this page, we offer extended pilot terms, direct founder access during the build, and a dedicated brand-voice training pass. Tell us during scoping.
Ecommerce Platforms
Helpdesk
Email / SMS
Reviews
Payment / Fraud
Returns
Communication
Inventory / Ops
Need a tool not listed? Tell us during scoping.
Brand voice is the make-or-break for ecommerce. We treat it as a first-class part of scoping. The process: you share 30-50 of your best customer-facing replies (support, marketing, social), your brand guidelines, and your tone notes. We train Patch and River on those. Before launch, you review sample drafts across 10-15 ticket types and flag anything that's off.
After launch, every reply is reviewable. We retune monthly during the first 90 days based on what your CX lead corrects. Brand voice gets sharper, not flatter, over time.
Each fellow is on a custom annual plan based on the role. No per-message charges. No credit meters. Predictable bills that scale with your brand, not your ticket count.
Default posture: replies are signed by a human (or by your brand without a personal name) and reviewed before sending. We do not impersonate specific employees. Some brands disclose AI use openly; others don't. We follow your call.
Volume scales without per-message charges. Annual pricing is predictable. During Black Friday or a viral moment, Patch handles the surge in the same workflow \u2014 ticket volume scaling is a feature of the model, not a billing event.
Default is draft-only. We can configure auto-process for narrow categories (returns within policy window, no friction signals) once you trust the patterns. Anything outside policy stays human-decided.
Helpdesk AI features run inside the helpdesk and apply vendor-default logic. Patch is custom-trained on YOUR brand voice and YOUR ticket patterns, lives in Slack so the whole team can ping it, and works across multiple channels (helpdesk + Instagram DMs + email) in one motion. Many brands run Gorgias's AI alongside Patch.
Tell us about your ticket volume. We will have Patch in your Slack in about a week.