From Microcontroller to AI Agent: Integrating Raspberry Pi Pico W with ASI Biont for Real-World Automation

Introduction

The Raspberry Pi Pico W is a $6 microcontroller that brings Wi-Fi connectivity to the RP2040 ecosystem. While it's powerful for standalone IoT projects—reading sensors, controlling relays, or blinking LEDs—its true potential unlocks when connected to an AI agent. ASI Biont, an AI-driven automation platform, can integrate with the Pico W using MQTT, SSH, or a COM port via Hardware Bridge. This article walks through a real-world case study: using a Pico W with a DHT22 temperature/humidity sensor and a relay-controlled fan, all managed through natural language commands in ASI Biont. No manual coding of the AI logic—just describe what you need, and the agent writes the integration code in seconds.

Why Connect a Pico W to an AI Agent?

Traditional IoT setups require you to write firmware for data collection, set up a cloud dashboard, and manually configure alerts. With ASI Biont, the AI agent becomes the brain: it reads sensor data via MQTT, analyzes trends, and triggers actions (like turning on a fan) when thresholds are exceeded. The Pico W acts as the physical interface—cost-effective, low-power, and Wi-Fi–enabled. According to the Raspberry Pi Foundation, the Pico W has sold over 4 million units since its 2022 launch, making it one of the most accessible microcontrollers for prototyping. Connecting it to an AI agent transforms it from a simple data logger into an intelligent edge node.

The Problem: Manual Oversight in a Small Greenhouse

A hobbyist runs a small indoor greenhouse with temperature-sensitive plants. The setup includes a Raspberry Pi Pico W connected to a DHT22 sensor (temperature and humidity) and a relay module controlling a 12V exhaust fan. The original workflow: the Pico W publishes sensor readings to an MQTT broker every 30 seconds, and the user manually checks a dashboard to decide when to turn the fan on. This resulted in missed alerts—the temperature spiked to 35°C (95°F) twice in one week, damaging seedlings. The user needed automated, intelligent control that could react to real-time data without requiring them to stare at a screen.

The Solution: ASI Biont + Raspberry Pi Pico W via MQTT

ASI Biont connects to the Pico W through MQTT, using the paho-mqtt library inside the execute_python sandbox. The user describes the setup in the chat: “Connect to my MQTT broker at 192.168.1.100:1883, subscribe to pico/sensor, and publish to pico/relay when temperature exceeds 30°C.” The AI agent writes a Python script that:

• Subscribes to the topic pico/sensor (where the Pico W publishes JSON like {"temp": 28.5, "hum": 60}).
• Parses the temperature value.
• If temp > 30, publishes ON to pico/relay; otherwise publishes OFF.
• Logs every reading to a local CSV file for trend analysis.

The Pico W firmware (MicroPython) is simple: it reads the DHT22 every 30 seconds and publishes via MQTT. The AI agent handles all decision-making and control logic.

Code Example: Pico W Firmware (MicroPython)

import network
import time
import json
from machine import Pin
import dht
from umqtt.simple import MQTTClient

# Wi-Fi credentials
SSID = "YourSSID"
PASSWORD = "YourPassword"

# MQTT broker
BROKER = "192.168.1.100"
TOPIC_SENSOR = b"pico/sensor"
TOPIC_RELAY = b"pico/relay"

# Sensor and relay
sensor = dht.DHT22(Pin(15))
relay = Pin(16, Pin.OUT)
relay.value(0)  # off initially

def connect_wifi():
    wlan = network.WLAN(network.STA_IF)
    wlan.active(True)
    wlan.connect(SSID, PASSWORD)
    while not wlan.isconnected():
        time.sleep(1)
    print("Wi-Fi connected")

def connect_mqtt():
    client = MQTTClient("pico", BROKER)
    client.connect()
    print("MQTT connected")
    return client

def read_sensor():
    sensor.measure()
    return sensor.temperature(), sensor.humidity()

def main():
    connect_wifi()
    client = connect_mqtt()
    while True:
        temp, hum = read_sensor()
        payload = json.dumps({"temp": temp, "hum": hum})
        client.publish(TOPIC_SENSOR, payload)
        print(f"Published: {payload}")
        time.sleep(30)

main()

Code Example: ASI Biont AI Agent Script (run in execute_python)

import paho.mqtt.client as mqtt
import json
import csv
from datetime import datetime

BROKER = "192.168.1.100"
TOPIC_SENSOR = "pico/sensor"
TOPIC_RELAY = "pico/relay"
TEMP_THRESHOLD = 30.0

csv_file = open("sensor_log.csv", "a", newline="")
csv_writer = csv.writer(csv_file)
csv_writer.writerow(["timestamp", "temp", "hum", "fan_state"])

def on_message(client, userdata, msg):
    data = json.loads(msg.payload.decode())
    temp = data["temp"]
    hum = data["hum"]
    timestamp = datetime.now().isoformat()

    if temp > TEMP_THRESHOLD:
        client.publish(TOPIC_RELAY, "ON")
        fan_state = "ON"
        print(f"Fan ON at {temp}°C")
    else:
        client.publish(TOPIC_RELAY, "OFF")
        fan_state = "OFF"

    csv_writer.writerow([timestamp, temp, hum, fan_state])
    csv_file.flush()

client = mqtt.Client()
client.on_message = on_message
client.connect(BROKER)
client.subscribe(TOPIC_SENSOR)
client.loop_forever()

Note: The sandbox has a 30-second timeout for execute_python, so for long-running MQTT loops, the user must run the script on their own server or PC. However, for quick tests, the AI can run a one-shot script that reads a single message and exits.

Results: What Improved?

The AI agent also provided a weekly summary: “Average temperature: 26.3°C, max: 33.1°C, fan ran 14% of the time.” This allowed the user to optimize the threshold to 28°C for better humidity control.

How the Integration Works: No Dashboard, Just Chat

To set this up, the user opens the ASI Biont chat and types: “I have a Raspberry Pi Pico W with a DHT22 sensor publishing to MQTT topic pico/sensor. I want an AI agent that subscribes to that topic, logs data, and turns on a relay via pico/relay when temperature exceeds 30°C. My MQTT broker is at 192.168.1.100:1883.” The AI responds with the Python script above, runs it in the sandbox, and starts monitoring. If the user wants to change the threshold, they just say: “Change the threshold to 28°C.” The AI edits the script and re-executes it. No buttons, no dashboards—everything happens through conversation.

Other Connection Methods for Pico W

While MQTT is the most natural for this Wi-Fi–equipped board, ASI Biont supports multiple protocols:

• SSH: If the Pico W is connected to a Raspberry Pi (via USB), the AI can SSH into the Pi and control the Pico W using minicom or screen.
• COM port via Hardware Bridge: For wired connections, run bridge.py on a PC connected to the Pico W via USB. The AI sends serial_write_and_read commands to toggle GPIO pins directly. Example: industrial_command(protocol="serial", command="serial_write_and_read", data="LED_ON\n").
• HTTP API: If the Pico W runs a simple web server (using microdot or tinyweb), the AI can use aiohttp to fetch sensor data or send commands.

Each method follows the same pattern: the user describes the connection parameters, and the AI generates the integration code.

The Bigger Picture: Why This Matters

The Raspberry Pi Pico W is just one of thousands of devices that ASI Biont can integrate with. Because the AI uses execute_python to write custom scripts, it supports any protocol that Python libraries can handle—Modbus, OPC-UA, CAN bus, BACnet, and more. This means a $6 microcontroller becomes a node in an intelligent, AI-driven automation system. For hobbyists, it eliminates the need to learn cloud APIs or complex orchestration tools. For professionals, it accelerates prototyping: describe the desired behavior, and the AI delivers a working integration in minutes.

Conclusion

Integrating the Raspberry Pi Pico W with ASI Biont transforms a simple IoT sensor into an intelligent edge device. The case study shows how an AI agent can monitor temperature, log data, and control a relay—all through natural language commands. The result: zero manual oversight, faster response to critical conditions, and actionable insights from logged data. Whether you're automating a greenhouse, a home lab, or an industrial prototype, ASI Biont lets you focus on outcomes, not wiring.

Ready to try it? Go to asibiont.com, describe your device and goal in the chat, and watch the AI write the integration in seconds. No dashboards, no delays—just your hardware, your AI.