No.226 AI 越强,创业者越要回到用户现场
Summary
This 三五环 episode has 刘飞 interviewing Kate, founder and CEO of AHA Creator, about applying AI to overseas creator marketing. It backfills Kate’s early operating path, AHA Creator’s 2024 product starting point, and the concrete workflow pain behind creator search, authenticity checks, multilingual outreach, negotiation, contracting, payment, and content review.
The durable contribution is a practitioner account of AI-Driven Creator Marketing as field-embedded workflow automation rather than AI product theater. Kate argues that AI is useful where work is repetitive but judgment-heavy: matching creators to products, reading audience authenticity, detecting PR risk, localizing communication, advancing collaboration, and performing first-pass brief checks, while customers still make budget, creative-direction, and final approval decisions.
Key Claims
- Kate says she began working on overseas growth after joining an early startup around 2021, when few domestic teams had repeatable web-product go-to-market experience for overseas markets.
- Creator marketing became salient because decentralized social media, KOLs, and KOCs could outperform conventional channels for some products; one YouTube creator case is said to have generated about 200,000 registrations after videos reached millions of views.
- The manual workflow is broad: teams search keywords, inspect channels and videos, judge creator fit, verify comments and engagement patterns, infer audience geography, collect email addresses, send low-reply cold outreach, negotiate, sign, pay, and review content.
- Kate argues that large models made creator matching more feasible than traditional algorithms because the data is incomplete and low-structure; judgment depends on reading content, style, past promotions, comments, and product-use context.
- AHA Creator uses AI for matching, fraud and PR-risk monitoring, multilingual communication, workflow follow-up, and first-pass content review against a customer brief.
- The product keeps human review around high-risk choices: campaign brief, creative angle, budget decisions, final creator selection, and final content approval remain with the customer.
- AHA Creator’s lower service-fee claim is tied to AI changing delivery cost: Kate contrasts traditional 30%-plus labor-heavy agency fees with AHA’s reported 5% service fee and token costs that may be below 1% of customer recharge.
- The platform is positioned as a two-sided product rather than a pure service business, with a brand-side workflow surface and a creator-side app for order taking.
- Customers reportedly like the “AI employee” form in Feishu groups because progress updates and approvals happen inside an existing collaboration channel instead of forcing users into a separate tool.
- Kate contrasts overseas creator marketing with many domestic placements: overseas campaigns are said to rely more on native creative, creator-product fit, and creator freedom, while domestic placements can resemble mechanically priced channel inventory.
Key Quotes
“AI 越强,创业者越要回到用户现场” - the source title’s product-building thesis.
“找达人其实就是一个非常需要人判断的事情” - Kate’s explanation of why creator matching was hard to automate with old software.
“最终这个钱还是客户花的” - Kate’s reason for keeping final creator and content decisions with customers.
Connections
- 三五环 and 刘飞 - show and host context for the interview.
- Kate and AHA Creator - central founder and company.
- AI-Driven Creator Marketing - core workflow concept extended by the episode.
- Overseas Creator Marketing Platform - two-sided platform and transaction-infrastructure category extended by the episode.
- Creator Trust Advertising / 创作者信任广告 - creator-product fit and native creative branch extended by the overseas comparison.
- AI Agent Overseas Commercialization, Agentic Workflow, IM Agent Interfaces, and AI Office Agent - adjacent AI workflow and interface frames touched by AHA’s AI employee and Feishu-group examples.
- TikTok, Instagram, YouTube, Xiaohongshu, Bilibili / 哔哩哔哩, Douyin, Feishu / 飞书, and DingTalk - platform and collaboration-tool context.
- Human Judgment Under AI, Automated Performance Marketing, and Global Product Localization - broader AI boundary, ad-automation, and localization concepts qualified by the source.
Contradictions
- No settled contradiction found with existing wiki content.
- The source reinforces and narrows the earlier 146 ✪ 对买量和铺货说bye-bye,AI如何驱动出海品牌增长? AHA Creator branch by adding more operating detail and a stronger human-in-the-loop product boundary.
- The transcript says it jumps from a 36:19 question about 2B AI product direction to English outro music after 36:56, so any missing later answer about SaaS, AS, MAS, or DingTalk remains unavailable from this source note.
- Service-fee, token-cost, creator-performance, customer-response, and matching-accuracy claims are founder-reported and should remain source-scoped pending independent evidence.