← Sticker Wall AGENT · DETAIL

Deploying AI Symptom Pre-screening Agents for SMB Telehealth Clinics, Generating $15,000/Month

Workflow: Every morning, import the self-reported symptoms and past medical records of the day's waiting patients into the triage

AGENT

Key Fields

FIELD STAMPS

IndustryHealthcare / Elderly Care

RegionUS（北美）

ScaleSME

ChannelOnline

🔧 Workflow

Every morning, import the self-reported symptoms and past medical records of the day's waiting patients into the triage agent. According to Curai's public case study standards, this HITL workflow can automatically complete about 90% of medical history collection and record organization. The system labels items one by one according to a three-tier red-yellow-green rule, where red alerts must be transferred to humans; it outputs triage suggestions, visit guidelines, and human-transfer work orders. Physicians only make final decisions on red and disputed cases, and new query methods are backfilled into the knowledge base weekly.

🛠 Setup Requirements

Requires the ability to use LLM APIs to build conversational workflows and perform basic compliance configurations. Tools include Claude or GPT-grade APIs, HIPAA-compliant customer service middle platforms (like Spruce Health), and API integration with the clinic's existing appointment system. Initial deployment for a single clinic takes about 2 to 4 weeks. A medical license is not required, but the clinic's licensed physician must review and fallback the dialogue scripts.

🧰 Toolchain

🔧 Claude API or GPT-4 grade model API
🔧 HIPAA-compliant communication middle platform (e.g., Spruce Health)
🔧 Clinic appointment system API
🔧 Conversation quality inspection and human-transfer work order dashboard

💰 Revenue

① SMB telehealth clinic monthly deployment subscription (main revenue): Each clinic pays $3,000/month. Serving 5 clinics = $15,000/month, equivalent to 100% of monthly revenue (unit price multiplied by the number of clinics is derived from the card data in the case study and has not been independently verified); ② Initial deployment implementation fee: New clinics pay a one-time fee of $5,000 to $10,000. The exact number of signed-up clinics has not been verified, and the proportion of this income in total revenue is blank; ③ Value-added subscription for compliance middle platform and human-transfer QC dashboard: Charge clinics an additional subscription fee for the middle platform and QC tools based on seats. The price per seat is not public, the number of purchased seats is unclear, and the proportion cannot be broken down; ④ Opportunity item — Commission per consultation volume: Referring to the Curai model of 24/7 online delivery by licensed physicians and practicing nurses with online prescription renewals (disclosed by the company itself), individual service providers can earn commissions based on the number of consultations. The commission rate is not stated, and the market share is unknown.

💸 Cost

Approximately $600 to $1,200 per month, mainly for LLM API call fees, compliant customer service middle platform subscriptions, and work order tool subscriptions.

⏱ Time Investment

2 to 3 hours per day during the operational phase for review and tuning; 4 to 6 hours per day for 2 to 4 consecutive weeks during the new clinic deployment phase.

🚀 Getting Started

Step 1: Do not take medical orders first. Use the same technology to build non-diagnostic appointment Q&A agents for dental clinics or physical examination institutions to practice and build up case studies; Step 2: Cold outreach to 3 small clinics in US telehealth entrepreneurial communities, trading a two-week free pilot for the first monthly fee contract, emphasizing in the pitch that AI only does pre-screening and physicians do the final review.

🔑 Keys to Success

✅ Strictly observe boundaries: AI only performs symptom collection and triage suggestions; diagnosis and prescriptions must be issued by licensed physicians. This is the compliance core of the Curai model.
✅ Turn human-transfer rates and mis-triage rates into weekly reports to show clinics, building renewal trust with data.
✅ Accumulate reusable triage knowledge base templates to cut the deployment time for the second clinic in half, creating compounding returns.
✅ Clearly state BAA agreements, data storage locations, and liability attribution in contracts. Ensure compliance before taking orders.

⚠️ 风险

⚠️ High medical compliance risks: If scripts cross the line to provide diagnoses, it may be deemed practicing medicine without a license. Contracts must explicitly state that physician final review liability rests with the clinic.
⚠️ Risk of patient health data leakage: Requires the use of compliant infrastructure and the signing of a BAA agreement.
⚠️ When misdiagnosis or missed triage causes disputes, technology vendors may be jointly held liable, requiring professional liability insurance.
⚠️ Major client concentration risk: Relying on a few clinics, the loss of a single client means the loss of thousands of dollars in monthly revenue.

📌 Real Cases

📌 Curai Health (San Francisco, founded in 2017): A virtual clinic featuring 24-hour AI pre-screening + online licensed physicians, having completed over 350,000 patient consultations with a 90% diagnostic accuracy rate. It won a global recognition award in 2026, proving the model's commercial and compliance feasibility.
📌 Hong Kong's 0xmd launched a free AI medical consultation bot in early 2026, supporting the upload of rash photos for preliminary interpretation, showing that lightweight AI health gateways are also gaining traction in Asia.
📌 IDC FutureScape 2026 Prediction: By 2030, 50% of Chinese graded hospitals will deploy medical AI agents to provide real-time decision support with over 80% accuracy and report abnormalities to clinical staff, meaning the AI pre-screening + human final review architecture is becoming the industry standard paradigm.

Sources