What if the history-taking was done before the appointment? InstantHPI makes that possible — at $0.003 per consultation, in 20+ languages, on any phone with Telegram.
Every clinic visit begins the same way: a doctor asks the patient what's wrong. The patient describes symptoms. The doctor asks follow-up questions. History is taken. Notes are written. That process takes 15 to 30 minutes — and it happens before any real medicine begins.
History-taking consumes 30–50% of every consultation. Doctors spend more time documenting than diagnosing. Documentation burnout is a leading cause of physician attrition worldwide.
Patients wait hours in overcrowded waiting rooms. In Nigeria, the doctor-to-patient ratio is 1 per 5,500 people. In Ethiopia, 1 per 10,000. Each consultation must count.
Oral histories get truncated. Key symptoms are omitted. Family history is never asked. The resulting clinical note is a pale shadow of what a structured intake would produce.
Most patients cannot name their medications, recall surgery dates, or articulate a clear chief complaint. The doctor wastes the first third of the visit extracting information the patient had but couldn't organize.
minutes spent per patient on history-taking and documentation
doctor-to-patient ratio in Nigeria
doctor-to-patient ratio in Ethiopia
cost of one AI pre-consultation that replaces that entire intake
The patient uses @instanthpibot on Telegram before their appointment. The bot guides them through a full medical intake — 18 questions, in their language, at their own pace. By the time they walk into the clinic, every document is already generated.
Opens @instanthpibot, selects language, answers 18 intake questions
7 clinical documents created in seconds, including differential diagnosis
Brings PDF printout or shows documents on phone to the doctor
Reviews pre-generated note, makes clinical decisions, signs where appropriate
More patients seen. Better notes. Reduced burnout. Zero extra cost.
Structured HPI with onset, duration, quality, severity, and modifying factors
Complete ROS across all organ systems — pre-populated from the intake
Past medical, surgical, social, and family history in structured format
Complete Subjective-Objective-Assessment-Plan — ready for physician review
AI-generated differential with clinical reasoning for each possibility
Suggested lab orders and imaging with clinical reasoning for each test
Pre-written referral letter for specialist or primary care handoff
The pre-visit consultation concept works across very different healthcare contexts. Here are four deployment models — each adapted to a different level of infrastructure, literacy, and physician involvement.
The patient does it on their own, unprompted
The clinic makes the bot part of its intake protocol
A CHW uses the bot on the patient's behalf
Patient uses the bot when symptoms first appear
Both sides of the consultation improve — at zero cost to either party.
At $0.003 per consultation, the numbers become almost impossible to argue with. This is not a pilot program that will cost millions to sustain — it is a service whose cost is essentially rounding error.
| Scale | Consultations / Year | AI Cost | Context |
|---|---|---|---|
| Single patient, 12 visits/year | 12 | $0.036 | Less than a single text message |
| Small clinic, 100 patients/month | 1,200 | $3.60 | Less than a coffee |
| Mid-size clinic, 5,000 patients/year | 5,000 | $15 | Less than one clinic supply order |
| 1,000 patients, monthly visits | 12,000 | $36 | Per year, for a community of 1,000 |
| Regional health system, 50,000 patients | 600,000 | $1,800 | Per year — less than one employee's monthly salary |
| National program, 1 million patients | 12,000,000 | $36,000 | Per year, for 1 million people with healthcare access |
The annual salary of a physician in sub-Saharan Africa averages approximately $15,000 USD. At $0.003 per consultation, that same $15,000 would fund 5 million AI pre-consultations — serving a population for which no physician currently exists. The question is no longer cost. It is reach.
We are looking for doctors, clinics, community health programs, and health systems to pilot these four models in real settings. Every environment is different. We want to learn from yours.
You already do everything yourself. Let the bot handle intake so you can focus on what only you can do — clinical judgment. Takes 5 minutes to pilot. Costs nothing.
High volume, constrained resources. Adding the bot to your check-in protocol is a zero-cost infrastructure upgrade. We can help train your reception staff.
Operating in underserved areas without reliable infrastructure? This model was designed for you. We want to understand your specific context and adapt accordingly.
Looking to extend primary care reach without new infrastructure investment? Let us walk you through the numbers and the deployment options. The technology is already live.
Interested in the triage model? We want data from real emergency settings — what urgency flags matter most, what the triage team would actually use, and how to reduce noise.
Every bot interaction is a complete clinical case with differential diagnosis and reasoning. A built-in teaching tool — for free, at any scale, in any language.
Tell us about your setting. We read every message personally — this is not a sales funnel. It's a conversation between people who care about the same problem.
Try the bot. Read the story. Partner with us. Help us understand how this fits your world.
cff@centremedicalfont.ca | Built by Dr. Carlos Font, Family Physician