---
title: "Cross-Border E-Commerce in the AI Era: From Factory Innovation to Customer Service"
author: "Unique Research"
sourcePublication: "Unique Research (FFCAP English library)"
originalPublishedAt: "2023-08-17T18:00:00+08:00"
canonical: "https://ffcap.cn/en/research/20230817a083je00"
source: "https://ffcap.cn/en/research/20230817a083je00"
language: "en"
cnSourceUrl: "https://view.inews.qq.com/a/20230817A083JE00"
---

# Cross-Border E-Commerce in the AI Era: From Factory Innovation to Customer Service

_Unique Research · Insight / AIGC Open Day · August 17, 2023 · Guests: QuickCEP, FOSHO, Attribuly, Unigravity (万有引力)_

2023: cross-border revenue up + domestic AI scale → AI×cross-border in focus. Intelligent drive via big data/ML for personalized products/services and predictive recommendations.

## QuickCEP — LLM AI shopping guide for China brands going global

Founded 2021: full-growth stack for D2C / native cross-border / known China brands—interaction, conversion, repurchase; smart guide + social tools; email/SMS/WhatsApp marketing automation. Hundreds of clients; focus growth-stage brands. Brand success patterns: unique positioning, shopping/after-sales experience, strong customer connection, data/tech-driven ops. LLM AI guide/CS: self-learn, direct business handling, global languages, reason/summarize. Three-layer LLM stack (foundation → industry corpus → scene prompts) + small models/flow bots for deterministic work; integrate site/social/WhatsApp. Advantages: no-train/no-config, 24/7, multilingual, reason/summarize, direct business, controllable cost; train on site catalog/brand story.

## FOSHO — AI affiliate marketing cloud

2021 Shenzhen (+HK/US): AI affiliate partner mgmt + sales closed loop; data-driven local channel ecosystems across localization funnel. AFF Copilot alliance sprite (competitive intel + channel recommend); AI Spark generative creatives (copy/image/video scripts) per channel. Head outdoor-brand case: insight→strategy→AI channel match → brand+performance lift.

## Attribuly — Shopify attribution & data intelligence

Automation for independent-site management—widely used in e-comm/cross-border for attribution and data ops.

## Unigravity (万有引力) — AI factory design

Import styles → redesign; AI can exceed human designers on logic/elements/materials; integrate craft models (carving/relief); stone/shape swaps; speed design/test loops. Open pattern + product libraries (~100k apparel/accessories); expand shoes/bags/home. AI across chain: efficiency, cost, custom print, platform partnerships; seek personalized valuable solutions for new e-comm forms.

## Roundtable

**Growth points:** Customs H1 cross-border trade ¥20T; online ~¥1T (export ~¥800B+)—online <5% of total → room. LATAM: Brazil (pop) + Mexico (near US, manufacturing); US: content commerce, DTC, offline→online; ID/ME infrastructure rising. Quality shift: red-ocean saturated; blue-ocean categories + multi-channel/services.

**Factory AI:** factories no longer only OEM—livestream/shop/ads; need low-barrier professional AI design/process tools fed by EU/US demand data.

**Customer AI use:** apparel e-comm active; brands strong at AI copy needing market edit; Amazon/DTC/SMB SEO & manuals; SMB copy quality especially.

**Will AI replace agencies?** Visual/brand-norm creatives still need human polish; agencies hold tech/data/overseas cognition and pick AI stacks; tacit vertical knowledge hard to fully automate.

**Landing barriers (Billy/QuickCEP):** selection→ads→txn→site→pay→consumer ops→maintenance all AI-able. Order: native ChatGPT/MJ → controllable paradigmatic image SaaS → 100% AI-rewritten steps. Ask additive vs multiplicative vs exponential per step.

**Vertical model future:** train OSS costly for most sellers → buy vertical capability from SaaS; each value-chain link may birth excellent vertical models; data-security consensus between client and vendor is the hard part. Domestic prefers private deploy; cross-border prefers tools.

_9 CN images; attach on ingest._

— END —

---

Editorial English edition published on UniqueCapital: https://ffcap.cn/en/research/20230817a083je00
On-site reading page: https://ffcap.cn/en/research/20230817a083je00