logo

Simple Agent

🏠︎

Simple Agent

×

Objective- Utilize AI to automatically generate E-prescriptions from text-based medical chats. Improve accuracy, streamline workflows, and ensure regulatory compliance with advanced AI-driven solutions.

How to Use- Here we are providing sample conversation between AI Agent(Doctor) and Patient for demonstration purpose.

Tools and Technology- Python, Langflow, OpenAI

Hello! I’m here to help gather your symptoms and medical history so I can prepare a Clinical Decision Support Summary (CDSS) for your doctor. Can you please describe what symptoms you're experiencing today, in your own words?
I’ve had a sore throat and slight fever since yesterday.
Got it — you mentioned a sore throat. How long has that been going on, and would you describe it as mild, moderate, or severe?
It’s been about a day. I’d say moderate.
Thank you. And you also mentioned a slight fever — do you know your temperature, or would you say it’s mild, moderate, or high?
It was 100.8°F last night. Still mild today.
Thanks for letting me know. Have you noticed any other symptoms?
Yeah, my voice is a little hoarse, and I feel tired.
Thanks for letting me know you’re also experiencing hoarseness and fatigue.
Now, a couple of quick questions about your medical history: Do you have high blood pressure (hypertension)?
Yes, I do.
And do you have diabetes?
No, I don’t.
Here’s what I have noted so far: - Symptoms: sore throat (moderate, 1 day), mild fever (100.8°F), hoarseness, fatigue - Chronic conditions: hypertension, no diabetes Is this summary correct? Please reply with a single word: yes or no.
Yes.
Thank you. Based on your symptoms and medical conditions, I have generated a Clinical Decision Support Summary (CDSS). Please consult a doctor for personalized care. (The JSON summary follows here.)

Output

Simple Agent

×

Team Members: 2–3 Members

Duration: 2 Months

Industry: Healthcare Industry

Technology: Python, Langflow

Introduction

The rise of AI conversational chatbots demands flexible, memory-aware assistants that can respond accurately and produce structured outputs. In this case study, we focused on building a smart chatbot in Langflow using OpenAI LLMs, chat memory support, and output formatting into structured JSON for integration with external systems.

The Challenge

  • Without memory, chatbots often feel disconnected or disjointed.
  • Lack of structured output that can be used by backend systems.
  • No simple way to control or customize prompts.
  • Limited ability to run custom logic alongside chat responses.

The goal was to build a simple, flexible chatbot that:

  • Remembers past conversations within a session.
  • Provides responses both for users and machines in JSON format.
  • Can be easily integrated with other systems through a UI or API.

The Solution

  • Chat Input: Captures real-time user input.
  • Memory Component: Stores session-based conversation history.
  • Prompt: Merges memory and current input for better contextual responses.
  • OpenAI Model: Connects via API to generate human-like replies.
  • Chat Output: Displays the response in the UI for user interaction.
  • Custom Component: Extracts and formats the AI output into JSON, and sends it to a specified URL.

Implementation

  1. Flow
    • Set up chat input and memory components.
    • Combine memory and live message into a unified prompt.
    • Connect the prompt to the OpenAI model.
    • Send model output to both the chat interface and the custom JSON component.
  2. Memory Component
    • Stores previous user messages during the session.
    • Feeds stored history into the prompt for context-aware generation.
  3. JSON Generator (Custom Component)
    • Extracts relevant information from OpenAI responses.
    • Formats output for use cases like order slips, reports, or logs.

Key Features

  • Maintains chat continuity through memory tracking.
  • Supports dual-output: user-friendly message and machine-readable JSON.
  • Prompt flexibility allows quick customization for different use cases.
  • Easily integrates with UIs and backends using APIs.

Results & Impact

  • Delivers more natural, memory-supported conversations.
  • JSON responses make it easy to connect with databases or third-party APIs.
  • Quick to deploy for those familiar with Langflow components.
  • Highly reusable for different applications like support agents, smart tools, or automation bots.

Conclusion

This simple agent shows how Langflow enables rapid chatbot development with memory support and structured outputs. The setup is ideal for use in customer support, internal tools, or smart applications. Langflow makes it easy to build, customize, and connect AI flows to real-world systems.