Unique Research · Insight / AIGC Online Open Day · October 10, 2023 · Full editorial rewrite of ~3,861 CJK — three company intros + Q&A preserved
Unique’s AIGC Online Open Day series (~2 sessions/month) hosts founders and executives on frontier AI and commercial opportunity. This session: AI × globalization — PMF signals, LLM product fusion, talent localization/cost, cross-border privacy/regulation, and 5–10 year industry change. Guests: Notta Inc, NovelMonkey, FlashIntel.
Companies
Notta Inc
Founded 2020, Shenzhen HQ — NLP and intelligent meeting assistants. SaaS speech-to-text: high-accuracy ASR + AI summary turns ~1 hour of audio into ~5 minutes of text with concise digests for interviews, in-person and online meetings. Web, desktop, iOS/Android, Chrome extension for browser audio. Features: real-time / file / Zoom·Meet·Teams·Webex auto transcription; tags/notes with audio–text replay; share links; export WAV/DOCX/TXT/PDF; AI summary; team workspace with permissions; encryption, access control, IP limits. 1.5M+ users; free + paid tiers. Japan market research cites clear burden/cost relief for transcription and meeting notes.
NovelMonkey
Founded 2021, Shanghai — AI original content. Core team ran Ctrip growth hacking; co-founded AnyBooks (overseas reading community, often #1 in 63 countries 2019–2020). >90% engineers; in-house Story LLM for 200k+ character mid/long novels; consumer products on that stack; multimodal content for multi-country users. Vision: “ByteDance of the AI era.” Author co-pilot: one idea → full novel for human revision.
FlashIntel
AGI-driven global GTM intelligence — products FlashInfo (revenue-growth platform to find/engage ideal B2B buyers) and FlashClick (pixel-to-SSL tracking, partnership insight). Team from ZoomInfo, Bombora, Salesforce, D&B, Equifax, etc. Founded 2022; 7 offices in 6 countries; 100+ engineers/scientists/GTM; 3,000+ global partners.
Roundtable
What marks PMF?
Chen Zhenwei @ Notta: Beyond need and value, willingness to pay is decisive.
Charles Tai @ NovelMonkey: Retention and sustained paying users. Definitions differ by investor — $100k revenue/profit in one market vs $1M elsewhere. Key: whether the company itself believes in sustained growth potential.
Shi Yi @ FlashIntel: Not only revenue scale — ROI, revenue efficiency/quality, LTV, and a living ROI model for growth path.
How LLMs enter product — future UX
Notta: LLMs collapsed barriers for STT, summary, action-item extract. Competition shifts to scene + data personalization from the user’s own data. AI will replace labor in many links and reshape product/business design — e.g. auto-summarize many/similar meetings, extract key points, machine orchestration for finer analysis.
NovelMonkey: Product is AI-generated; started from product pain + Silicon Valley inspiration; iterated via knowledge nets and retrieval into an AI tech framework for changing demand.
FlashIntel: AI value = data. ToC: DAU/MAU; ToB: sector/scene knowledge + data enablement. Algorithms/compute matter less when GPT APIs are open. Use cases: lower CS cost, search, recommendations, email/scripts, STT, sentiment. Choose AI-assist vs AI-native by need.
Talent localization & cost
Notta: Japan expansion needed local sales/CS; language and culture still hard for AI to replace for customer intimacy.
FlashIntel: Sales/growth on-site; other roles remote. Post-COVID remote org across time zones needs quantified outcomes — easier for GTM than for eng (commits, quality, issue resolution).
Data privacy & regulation
Notta: Platforms like Vanta.com for GDPR/ISO27001 cheaper than law firms for SMBs; give users controls to limit enterprise data into GPT training.
NovelMonkey: Store/platform listing requirements on data use are usually clear; AI regulation drafts (e.g. UK from Mar) still early — stay careful.
FlashIntel: Heavy compliance by country IP; GDPR and EU-region storage; collect/opt-out transparency. SaaS compliance tools often beat expensive law-firm outsourcing.
5–10 year change
Notta: Repetitive work → AI; human creativity/imagination gain time.
NovelMonkey: AI becomes infrastructure — humans + AI co-generate and co-consume content in an invisible data loop; humans as feedbackers.
FlashIntel: Lower cognitive cost for white-collar think/write/plan work; ML beyond LLMs still advances; GPT-5/6/7 arrive faster; emphasis shifts to service quality in hospitality and similar.