AI-Powered Inventory Cycle Count Bot via Telegram and Google Sheets - n8n Workflow

• Warehouse managers looking to digitize cycle counting.


• Logistics experts needing efficient, voice-activated inventory solutions.


• Developers seeking advanced examples of n8n templates integrating messaging apps and AI.


• Companies needing to improve data accuracy in their cycle count processes using an n8n solution.

This sophisticated n8n workflow transforms manual, paper-based inventory counting into an efficient, voice-activated digital process. It uses a Telegram n8n trigger to capture voice commands from warehouse operators, leverages OpenAI's AI models to accurately transcribe and extract structured data (location ID and quantity), and instantly updates a Google Sheet.

This specific n8n workflow ensures data accuracy by validating the extracted information against a master list of locations before committing the update. It’s an ideal example of how specialized n8n templates can handle complex, real-world logistics challenges, offering a robust, low-code automation solution built entirely on n8n.

1. Trigger & Routing: The process starts with the Telegram Trigger n8n node when an operator sends a message. The flow uses an If n8n node to determine if the input is a text command (like /start or /help) or a voice message.


2. Command Execution: If the command is /start, the n8n workflow fetches all remaining, unchecked locations from Google Sheets using a specialized Google Sheets n8n node and instructs the operator on the next location to count using another Telegram n8n node.


3. Transcription & Extraction: For voice messages, the Collect Audio n8n node downloads the file. The Transcribe Operator Command n8n node (OpenAI) converts speech to text. A subsequent AI agent, utilizing the Get Location, Qty n8n node, then employs an LLM and a Structured Output Parser to reliably extract the critical inventory data (locationid and quantity).


4. Validation: The Transcription Error n8n node checks if the AI successfully extracted both required values. If successful, the system performs a critical check via the Is Location in Scope? n8n node to verify if the reported location is valid and currently slated for counting. If validation fails at any point, the operator receives an appropriate error message via Telegram.


5. Data Update: If validated, the Update Inventory Quantity Google Sheets n8n node updates the row corresponding to the locationid with the actual quantity counted. The n8n workflow then initiates the instruction for the next item to be counted, ensuring a smooth cycle count flow.

1. Import the n8n Workflow: Download this n8n workflow JSON file and import it directly into your self-hosted or cloud n8n instance.


2. Credential Setup: You must configure credentials for three services:

1. Google Sheet Structure: Ensure your primary inventory spreadsheet (referenced in the Google Sheets n8n nodes) contains mandatory columns: locationid, systemquantity, actual_quantity, and checked (used to mark counted locations, typically 'X').


2. Activate the n8n Trigger: Activate the workflow to register the webhook with your Telegram bot, allowing the Telegram Trigger to begin receiving messages.

Telegram Trigger: This is the initial n8n trigger node that starts the cycle count process upon receiving any message from the configured Telegram chat.
Collect Audio (Telegram node): An essential n8n node for the voice-based process, responsible for downloading the binary file of the operator's voice message.
Transcribe Operator Command (OpenAI node): Utilizes OpenAI's speech-to-text capabilities to convert the audio input into usable text data for downstream processing by the n8n workflow.
Get Location, Qty (Langchain Agent node): A powerful AI n8n node configured with system prompts to reliably extract structured data (location ID and quantity) from the potentially messy transcribed text, crucial for successful data update.
Structured Output Parser (Langchain node): Ensures that the LLM output conforms strictly to the JSON schema defined for the inventory data.
Google Sheets n8n nodes: Used in two key roles: retrieving the list of outstanding locations (Collect Remaining Locations and Get Locations to Check) and committing the final count data (Update Inventory Quantity).

• If n8n nodes (e.g., Transcription Error, Is Location in Scope?): These core logic n8n nodes control the flow, ensuring that if data extraction fails or the location is invalid, the operator receives immediate feedback instead of updating incorrect records.