AI Telehealth Chatbot PoC and Investor Pitch Deck Delivered in Two Weeks

Summary

In two weeks, ScienceSoft delivered a PoC for an AI chatbot meant to streamline patient diagnostics before telehealth sessions and provide clinical decision support to physicians. The client also received clickable UI prototypes and an investor pitch deck.

About Our Client

The Client is an innovative US startup aiming to transform telemedicine with AI solutions that improve diagnostics, automate patient triage, and empower healthcare providers with robust clinical decision support.

Trusting ScienceSoft’s experience in developing AI and medical software, including intelligent chatbots for healthcare, the Client turned to the company to create a proof of concept (PoC) for its AI-powered chatbot. The chatbot was meant to collect and analyze patient-reported symptoms, offer preliminary diagnostics, and recommend next steps, such as scheduling telehealth appointments, seeking emergency care, or managing conditions at home. Additionally, the chatbot was intended to suggest over-the-counter medications and lifestyle adjustments to patients, generate tailored reports for patients and physicians, and provide doctors with clinical decision support. The solution was to be integrated with telehealth platforms, popular EHRs, drug databases, and insurance services. The Client needed the bot to be fully compliant with HIPAA, GDPR, and WCAG standards for virtual care delivery.

To attract funding and lay the groundwork for future development, the startup also needed a compelling demo, including clickable UI prototypes and an engaging investor pitch deck.

2 Weeks and 4 Talents to Deliver AI Chatbot PoC, Clickable Prototypes, and Investor Pitch Deck

To complete the project within a short timeframe, ScienceSoft assembled a team of a business analyst, UI designer, data scientist, and front-end developer. The team analyzed the startup’s requirements for the future solution and chose to build the AI chatbot PoC using Azure Health Bot, as it already had a strong foundation for regulatory compliance and several essential functional modules, including symptom checking, a clinical knowledge base, appointment scheduling, analytics, and reporting.

The delivered PoC had the following patient-facing functions:

• Symptom input. Patients describe their symptoms via a web-based interface, answering questions posed by the chatbot. If needed, they can select an option from an AI-predicted list of responses.
• Preliminary diagnosis. Based on the reported symptoms, patients receive preliminary diagnoses in layman’s terms. Depending on the severity of their condition as determined by the bot, patients receive recommendations for the next steps, such as self-monitoring, seeking emergency care, or scheduling a doctor’s appointment. The bot suggests a list of available providers from integrated telehealth platforms (e.g., Teladoc Health, Amwell, Doctor on Demand, MDLive, Doxy.me) for virtual care delivery.
• Treatment recommendations. For non-urgent conditions, patients get over-the-counter drug recommendations (e.g., acetaminophen), including usage instructions and potential side effects. They are also provided with educational content, such as symptom explanations, general tips on lifestyle adjustments and preventative measures, as well as FAQs.
• Reports. Patients receive chatbot-generated reports detailing potential conditions and recommendations on their management. These reports can be rendered as PDFs in real time for sharing, downloading, or printing.

The physician-facing features included:

• Clinical decision support. Via an integrated dashboard, doctors receive reports generated for each patient. These include patient medical history (including lab test results), information on the defined triage level, symptom records, diagnostic results (including differential diagnoses), suggested treatment options (including first-line treatments, alternative therapies, and potential drug interactions), and interdisciplinary care notes. Physicians can ask patients follow-up questions for better context and decision-making. The bot uses medical terminology and evidence-based references to clinical guidelines, research studies, and other sources to support the provided information.
• Report modification. Physicians can customize treatment plans by adjusting parameters like dosage and treatment duration in the chatbot-generated reports.

During the proof-of-concept stage, the chatbot operated in anonymous mode and did not collect users’ PHI. The Client’s future plans included allowing users to create secure patient profiles to store their clinical and non-clinical information.

In addition, the startup aimed to integrate the solution with popular EHRs (Epic, Cerner, Athenahealth, Allscripts, and eClinicalWorks), drug databases (Micromedex, DrugBank), and insurance services.

For the demo, ScienceSoft’s UI designer customized the chatbot’s standard interfaces to align with the Client’s branding and created clickable prototypes showcasing the bot’s functionality.

To support the startup’s investment efforts, ScienceSoft’s business analyst prepared a pitch deck highlighting the AI chatbot's potential and market viability. This presentation was crafted to effectively communicate the financial and operational benefits to prospective investors.

With the PoC and demo phase complete, the Client was ready to pursue fundraising to advance the chatbot’s MVP development.

Key Outcomes for the Client

• A stronger foundation for attracting MVP funding. In just two weeks, ScienceSoft delivered a functional AI chatbot PoC, clickable UI prototypes, and an investor pitch deck, enabling the startup to demonstrate its product concept and present its business case to potential investors.
• A tangible product concept to guide MVP development. The chatbot PoC showcased patient triage, preliminary diagnostics, and clinical decision support, giving the Client a practical starting point for planning the next development stage.
• Reduced upfront development effort. By building on Azure Health Bot's existing capabilities, ScienceSoft helped the startup reach the demo stage without developing the chatbot's core functionality from scratch.

Technologies and Tools

Azure Health Bot, Figma, PowerPoint, HTML, CSS, JavaScript, Python Flask.