← Sticker Wall AGENT · DETAIL

Second-Level Public Opinion Radar and Crisis Statement Draft All-in-One Subscription: Corporate Annual Fee Starting from 120,000 RMB

Workflow: Every day, brand-related text, short videos, and social platform sources are crawled across all media via the 乐思 Public

AGENT

Key Fields

FIELD STAMPS

IndustryMarketing / Advertising

RegionChina（中国大陆）

ScaleSME

ChannelOnline

🔧 Workflow

Every day, brand-related text, short videos, and social platform sources are crawled across all media via the 乐思 Public Opinion API and 识微 Agent interface. A fine-tuned sentiment analysis model is used to perform second-level anomaly scoring on negative signals, focusing on three abnormal scenarios: sudden surges in negativity, skyrocketing forwarding volumes, and mentions by high-weight accounts (such as official media and top KOLs). Once preset thresholds are triggered, the LLM API is automatically called to generate 3 versions of PR statement drafts with different perspectives (including fact-clarification, responsibility-taking, and apology-rectification versions), which are simultaneously pushed to the client's designated contact via WeChat Work for review. Once confirmed by the client, they can be distributed to official channels with one click.

🛠 Setup Requirements

There is no need to build a whole-network data collection system from scratch. Directly subscribe to the 乐思 Public Opinion Monitoring commercial API or 识微 Business Intelligence Agent to obtain standardized public opinion data streams, and procure the Claude API or Tongyi Qianwen LLM API for statement draft generation. Basic proficiency in using automation tools like n8n is required to connect the full-link workflow of data cleaning, anomaly detection, draft generation, and message pushing, along with configuring WeChat Work group bots for hierarchical early-warning pushes. Building the overall system from scratch takes about 4 weeks, with no complex coding development required.

🧰 Toolchain

🔧 乐思 Public Opinion Monitoring SaaS
🔧 识微 Business Intelligence Agent
🔧 Claude API
🔧 WeChat Work Group Bot
🔧 n8n Automation Workflow

💰 Revenue

① Integrated subscription annual fee (main revenue): Brand owners pay 120,000 to 200,000 RMB annually (including monitoring + statement drafts + quarterly reviews). A single operator serves 10-15 clients, generating annual revenue of 1.2 million - 3 million RMB, with a monthly average of about 100,000 - 250,000 RMB (the proportion of this path in total revenue is not specified, case statement, independent review absent, as of 2026); ② Tiered subscription: Basic monitoring annual fee of 30,000 RMB, statement draft generation only annual fee of 80,000 RMB, tiered according to client needs. The exact number of contracts signed for each tier cannot be verified (case self-description, external unverified), and the proportion after tiering is not listed; ③ Capability white-label resale: Authorizing and reselling second-level monitoring capabilities. The source claims 7×24 hour coverage of 20+ platform types with 5-minute updates (merchant caliber, without third-party verification). Resale pricing is not public, and there are no figures on how much revenue this part can contribute; ④ Opportunity item - Early warning premium tier: The source claims AI can discover crisis clues 2 hours in advance, whereas traditional monitoring takes 2-4 hours (merchant self-description, independent validation absent), which can be used as a high-premium tier, with its proportion currently vacant.

💸 Cost

The 乐思 Public Opinion commercial API annual fee is about 20,000-30,000 RMB, the 识微 Business Intelligence Agent basic version annual fee is about 15,000 RMB, the LLM API monthly consumption is about 3,000 RMB (calculated based on generating 500 statement drafts per month), and the monthly cost of automation tools (n8n cloud version + server) is about 1,000 RMB. Total monthly operating costs can be controlled between 8,000 RMB and 12,000 RMB.

⏱ Time Investment

Invest 3 hours daily in early warning reviews, statement draft polishing, and client communication, about 20 hours per week, totaling about 80 hours per month. Single-client service duration is about 8 hours per month, with the remaining time allocated to system iteration and new client acquisition.

🚀 Getting Started

Novices start with the 14-day free trial of 识微 Business Intelligence or 乐思, familiarize themselves with public opinion warning rules and API call logic, and find 1-2 local SME clients with brand public opinion pain points (such as chain catering or regional retailers) to conduct a 1-month trial run of manual monitoring plus AI drafts, producing weekly reports including response timeliness comparisons to prove value. After running the minimum viable loop, sign contracts at an annual fee of 120,000 RMB, and subsequently replicate the same system to batch-serve clients in the same industry.

🔑 Keys to Success

✅ Second-level anomaly detection threshold tuning capability determines early warning accuracy
✅ Statement drafts must provide multi-version perspectives for legal selection
✅ Annual subscriptions bind clients to long-term trust
✅ Full-process automation from early warning to drafting compresses response time

⚠️ 风险

⚠️ An excessively high false-positive rate will desensitize clients to system alerts, causing real crises to be ignored when they occur and resulting in a loss of client trust
⚠️ If the AI-generated PR statement draft does not meet brand legal compliance requirements or conflicts with the official stance, it may trigger secondary public opinion risks, requiring strict manual review
⚠️ Enterprise clients have extremely high requirements for public opinion data security; if the collection and storage links fail to pass Multi-Level Protection Scheme (MLPS) Level 3 certification, signing medium and large clients will be hindered

📌 Real Cases

📌 The 识微 Business Intelligence Agent has been explicitly showcased in iResearch's official introduction for its all-media negative monitoring and automated report generation capabilities, serving over 2,000 enterprise clients with a mature annual fee model
📌 The 乐思 Crisis PR Early Warning Solution targets government, enterprise, and brand clients, charging by the year with support for minute-level WeChat pushes, and single-client annual fees reaching up to 150,000 RMB
📌 A regional chain tea brand in East China (with annual revenue of about 400 million RMB) integrated a similar second-level public opinion radar system in 2025, compressing the average response time for negative events from 7.5 hours to 25 minutes, reducing annual crisis PR disposal costs by 34%, and achieving a 100% annual renewal rate

Sources