VerifyScript

VerifyScript is a web-based AI application that helps patients, caregivers, and providers analyze the safety and effectiveness of multiple medications. By combining structured clinical data with GPT-4, VerifyScript generates clear, actionable medication assessment reports, reducing confusion, uncovering risks, and supporting safer care conversations.

Highlights

• AI-generated medication review reports in minutes
• Supports prescription, OTC, and supplement analysis
• GPT-4 + RxNorm clinical grounding
• Stripe-gated, subscription-based assessments
• Built and launched in 6 weeks

Quick Facts

• Industry: Healthcare / Digital Health
• Platform: Web Application
• Users: Patients, caregivers, providers
• Engagement Model: B2C, subscription per assessment
• Stackup Services: Product strategy, full-stack development, AI integration

Background

Medication regimens are increasingly complex, especially for older adults managing multiple conditions. Patients often leave appointments with prescriptions they don’t fully understand, struggle to read medication labels, or remain unaware of interactions between drugs, supplements, and food.

VerifyScript was created to address a simple but dangerous gap: patients do not have an accessible way to understand the real-world consequences of taking multiple medications together.

The goal was to build an MVP that could:

• Translate complex medication data into plain language
• Surface interaction risks and prescribing cascades
• Equip patients with a report they can confidently share with their provider

The Challenge

Business Challenges

• Medication safety tools are often designed for clinicians, not patients
• Existing solutions are fragmented across pharmacists, PDFs, and web searches
• Patients lack a single, understandable source of truth
• The product needed to prove willingness to pay quickly

Technical Challenges

• Converting messy medication inputs (photos, text, voice) into structured data
• Grounding AI outputs in real clinical terminology
• Avoiding unsafe medical advice while still providing meaningful insight
• Generating professional, shareable reports automatically

Operational Constraints

• Pre-HIPAA MVP with clear legal boundaries
• No internal admin dashboard—operations had to stay lightweight
• A strict 6-week delivery timeline

The Solution

Stackup Solutions partnered with VerifyScript to design and build a focused, production-ready MVP optimized for speed, safety, and validation.

Product Architecture

• Frontend: React.js web application
• Backend: Node.js + Express
• AI Layer: OpenAI GPT-4
• Clinical Data: RxNorm for medication normalization and auto-complete
• OCR: Google Vision API with post-processing cleanup
• Payments: Stripe (required before AI execution)
• Reporting: HTML-to-PDF generation using Puppeteer
• Email Automation: Gmail API

Core Features Delivered

1. Medication Intake Engine

• Manual entry for flexibility
• OCR photo upload for prescription labels
• Cleaned, normalized inputs mapped via RxNorm
• Auto-complete across brand, generic, and drug class

2. GPT-Powered Medication Analysis

The AI engine analyzes submitted medications to identify:

• Drug–drug interactions
• Drug–food interactions (enhancing or reducing effect)
• Prescribing cascades (meds treating side effects of other meds)
• Deprescribing opportunities (flagged for provider discussion only)
• Monitoring markers patients should be aware of
• Recent recall notices

All outputs are educational, never directive, and consistently prompt users to consult a healthcare professional.

3. Structured Safety Guardrails

• Anonymized patient inputs only
• Prompt constraints prevent definitive medical instructions
• Clear disclaimers embedded throughout the report
• Risk insights framed as discussion points, not decisions

4. Branded PDF Report Generation

• AI-generated content compiled into a professional PDF
• Clear sections, summaries, and next steps
• Designed to be shared directly with providers

5. Monetization & Automation

• Stripe payment required before analysis runs
• Subscription-based per assessment
• Automated report delivery via email
• Follow-up thank-you email sent after 3 days

Implementation Process

Phase 1: Product & Risk Definition

• Defined MVP scope to stay outside FDA SaMD territory
• Established AI boundaries and disclaimer strategy
• Designed the assessment-to-payment flow for conversion

Phase 2: Core Build

• Built medication intake and OCR pipeline
• Integrated RxNorm for structured medication intelligence
• Developed GPT-4 prompt framework for safe, sectioned outputs

Phase 3: Reporting & Monetization

• Implemented Stripe gating
• Built backend PDF generation pipeline
• Automated email delivery and admin tracking

Despite the compressed timeline, Stackup delivered a fully functioning, monetized MVP ready for real users.

Results & Early Impact

As an early-stage MVP, VerifyScript focused on validation over scale.

Initial Outcomes

• Successfully launched within the 6-week timeline
• First paid users onboarded within the initial release window
• Achieved early traction toward the goal of 10+ paid users in 90 days
• Validated willingness to pay for AI-generated medication assessments
• Demonstrated strong user interest in interaction and prescribing cascade insights

Qualitative Signals

• Users reported discovering interactions they were previously unaware of
• Reports helped patients feel more confident discussing medications with providers
• Caregivers valued having a single, understandable summary