← Sticker Wall AGENT · DETAIL

Financing Roadshow Recording to IR Whitepaper: A Solo Consultant Making 24k RMB Monthly

Workflow: Every morning, I first sync the list of pending recordings, prioritizing projects to be delivered within 48 hours. The i

AGENT

Key Fields

FIELD STAMPS

IndustryFintech

RegionChina（中国大陆）

ScaleSME

ChannelOnline

🔧 Workflow

Every morning, I first sync the list of pending recordings, prioritizing projects to be delivered within 48 hours. The inputs include the client's roadshow video/audio files, basic financial data packages, and track background descriptions. The first step uses Whisper-like tools for high-precision transcription. The second step uses large language models to extract core data points, business model keywords, and competitor benchmarking information. The third step generates the first draft according to the fixed structure of the IR whitepaper. After manual cross-verification of data and terminology polishing, the PDF delivery is outputted, while the template library for that track is simultaneously updated.

🛠 Setup Requirements

Requires a Windows/Mac computer with 16GB RAM or more, basic prompt engineering skills, and an understanding of the primary market's standard financial statements structure. No programming background is required. The setup period takes about 10 days: the first 3 days test the accuracy of transcription tools, the middle 5 days build IR whitepaper templates for 3 different tracks, and the last 2 days onboard the first paid verification client. In the early stage, the local deployment solution of the open-source project AI-Media2Doc can be used to lower tool costs.

🧰 Toolchain

🔧 BibiGPT
🔧 OpenAI Whisper
🔧 Sonix
🔧 NotebookLM
🔧 Notion

💰 Revenue

① Startup project basic projects (main revenue): Clients pay a service fee per project. Basic version roadshow to IR whitepaper is 8,000 RMB/order × 3-4 orders/month = 24,000-32,000 RMB/month, which matches Kane's current monthly income of 24,000 RMB, accounting for about 100% of the monthly income (estimated figure, case caliber, independent review not done); ② Annual service package repurchases: Clients subscribe annually to service packages including quarterly data updates, 36,000 RMB/order × 1-2 annual clients = 36,000-72,000 RMB/year (monthly average 3,000-6,000 RMB, estimated), which is the incremental factor pushing monthly income above 40,000 RMB/month. Its exact share is not specified (case statement, independent review missing); ③ FA and venture capital financial advisor channels: Paying rebates or joint delivery per closed deal to channels. The exact rebate amount is not public, and the number of processed deals has not been verified, leaving the proportion of total revenue blank; ④ Opportunity item - Overseas Pre-A project English IR whitepaper: Charged per project. Benchmarked against the manual transcription market rate of 1-2 USD per minute, taking 24-48 hours (prices taken from platform public pricing), AI delivery can be compressed to 1 day. The exact revenue share here remains unquantified.

💸 Cost

Transcription tools are billed by recording duration at about 5-10 RMB per hour. AI generation API monthly cost is about 300-800 RMB. Collaboration tools like Notion team edition cost about 100 RMB per month. Overall monthly tool costs are controlled within 1,500 RMB with no high fixed subscriptions. Marginal costs remain almost unchanged as order volume increases.

⏱ Time Investment

Dedicate about 15 hours per week. A single project takes about 3 working days from receiving the recording to delivery, with transcription and AI generation taking 1 day, and manual data verification, terminology polishing, and aligning caliber with the client's CFO taking 2 days. Delivery schedules can be staggered when handling multiple concurrent orders.

🚀 Getting Started

Step 1: Find 3 public early-stage project financing roadshow videos, run them through transcription tools and large models to produce sample IR whitepapers for 3 different tracks (e.g., hard tech, consumer, enterprise services), focusing on polishing the accuracy of three modules: financial data extraction, competitor benchmarking, and risk disclosure. Step 2: Proactively contact venture capital financial advisors and small/medium project teams on FA platforms, offering 1 sample for free as proof of capability. After securing the first order, quickly build up the exclusive template library for that track, allowing subsequent similar projects to be replicated in batches.

🔑 Keys to Success

✅ Compounding effect of the industry financial template library: Every completed track project accumulates a set of comparable company metrics, industry financial terminology libraries, and standard disclosure structures, increasing efficiency for subsequent similar projects by over 60%.
✅ Human CFO acting as data referee: AI is only responsible for generating the first draft. All financial data and revenue recognition calibers must be cross-verified by a licensed CFO, which is the core barrier distinguishing this from pure transcription services.
✅ Value-based pricing per delivery rather than hourly billing: Lock in referral channels from FAs and venture capital financial advisors to avoid falling into price-war transcription traps.
✅ Extend the service chain from one-off transcription to high-margin value-added services such as mid-term quarterly data updates, due diligence Q&A packages, and investor communication script generation, thereby increasing client LTV.

⚠️ 风险

⚠️ Pure transcription services have extremely low margins; without building industry financial knowledge barriers, they will be directly replaced by big tech meeting minutes products and low-cost transcription outsourcing, losing pricing power.
⚠️ Client financial data leakage risk: Financing roadshows contain a large amount of undisclosed revenue and user data. Without comprehensive NDAs and data security measures, any leakage will result in heavy compensation and reputational damage.
⚠️ Unstable accuracy of AI-generated content: Errors in financial data extraction or non-compliant disclosures could hinder the client's IPO process, triggering legal disputes and customer complaints.

📌 Real Cases

📌 BibiGPT's public 'meeting video to structured report' workflow has been verified by tens of thousands of workplace users. A single person using AI tools can convert a 2-hour roadshow recording into a 30-page IR whitepaper in just 1 day, which is over 5 times more efficient than manual writing.
📌 Sesame Pie AI reported an enterprise content service provider using NotebookLM to convert enterprise existing documents into structured podcasts, whitepapers, and other derivative content, achieving a monthly income of 12,000 RMB, proving that there is a clear willingness to pay for enterprise existing content recreation services.
📌 An AI roadshow material generation tool tested by a domestic VC FA in 2025 processed 27 roadshow recordings in a single month, compressing the manual writing cycle from the original 3 days to 4 hours. The project owners' payment rate increased from 32% to 67%, validating B-end willingness to pay.

Sources