Why construct issues the exhausting method when you possibly can design them the sensible method?
As a Provide Chain Knowledge Scientist, I’ve explored varied frameworks like LangChain and LangGraph to construct AI brokers utilizing Python.
The illustration above is from an article I wrote on the finish of 2023, titled “Leveraging LLMs with LangChain for Provide Chain Analytics — A Management Tower Powered by GPT.”
On the time, I used to be exploring the way to use LangChain to construct an agent performing as a Provide Chain Management Tower.
A yr later, I found the facility of the low-code platform n8n to construct the identical sort of resolution in just some clicks.
On this article, we’ll discover the way to simply construct AI brokers to automate provide chain analytics workflows utilizing n8n.
We’ll additionally see the way to deploy the identical AI-powered Management Tower agent I initially constructed with LangChain 18 months in the past — now utilizing solely low-code.
AI Agent for Provide Chain Management Towers utilizing LangChain
My first undertaking of AI Automation undertaking utilizing n8n was for a buyer who needed a Provide Chain Management Tower outfitted with a chat interface.
A Provide Chain Management Tower is a set of dashboards and reviews related to Warehouse and Transport Administration Programs that use knowledge to watch essential occasions throughout the availability chain.
In an earlier article revealed on In direction of Knowledge Science, I experimented with LangChain to attach a management tower to an AI agent.
The thought was to construct a plan-and-execute agent that will
- Course of the consumer’s request written in plain English
- Generate the suitable SQL question
- Question the database and retailer the outcomes
- Formulate a transparent response in plain English
After a number of iterations, I discovered the precise chain construction and prompts to ship correct outcomes.
The answer labored effectively as a result of I had already gained expertise utilizing LangChain and different frameworks to construct AI brokers.
How are we supposed to take care of this advanced setup?
Nonetheless, to supply this as a service, I wanted instruments that will make the answer simpler to deploy, keep, and enhance — even with restricted Python information.
That’s after I found n8n.
Let’s dive into that within the subsequent part.
AI Agent for Provide Chain Management Towers — Constructed with n8n
What’s n8n?
n8n is an open-source workflow automation software that allows you to simply join apps (e-mail, CRMs, messaging methods), APIs, and AI mannequin frameworks like LangChain.
You construct workflows by connecting pre-built nodes.
As an example, the workflow above processes emails
- The primary node collects emails from a Gmail account.
- The e-mail content material and metadata are despatched to the AI Agent node, which extracts the related info.
- The third node processes the output utilizing JavaScript.
- The ultimate node masses the outcomes right into a Google Sheet.
No code was wanted to construct this workflow — apart from the third node, which makes use of simply two strains of JavaScript.
Since I work with a group of Provide Chain consultants who’ve restricted Python expertise, this was a game-changer for me as I appeared to develop my service providing.
They will simply use, adapt, and keep this workflow after a brief coaching session on n8n.
AI Provide Chain Management Tower n8n workflow
The AI Provide Chain Management Tower workflow is a little more advanced — however nonetheless far less complicated than its Python model.
It consists of two sub-workflows.
The principle sub-workflow consists of each a chat interface and the AI agent.
For the AI Agent node, you want to
- Join an LLM (chat mannequin) utilizing a node the place you enter your API credentials
- Add a reminiscence node to handle the dialog
- Add a software node for SQL querying, linked to the second sub-workflow
The AI agent generates an SQL question and sends it to the “Name Question Instrument” node, which executes the question.
The sub-workflow features a code node that cleans the question (eradicating further areas and blocking dangerous instructions like DELETE).
The output is shipped to a BigQuery node, which runs the question and returns the outcomes.
The method may be very clean and requires restricted configuration:
- System Immediate (within the AI Agent node)
- Person Immediate (within the AI Agent Node)
This setup requires no Python expertise and will be dealt with instantly by my consultants.
The outcomes are akin to these of the Python model.
For step-by-step setup directions, take a look at my YouTube tutorial 👇
Conclusion
This instance exhibits how simple it’s to copy an AI agent constructed with Python — utilizing n8n and minimal code.
Does that imply Python is now not wanted for Provide Chain Analytics? Undoubtedly not!
Like many low-code platforms, the options are restricted to what’s accessible inside the framework.
That’s why I take advantage of it as a complement to my analytics merchandise.
To try this, you should utilize the HTTP Request node to attach your workflow to your analytics backend.
What else? Straightforward connectivity to many companies.
One more reason I selected n8n to counterpoint my analytics merchandise is how simple it’s so as to add further connections.
For instance, if you wish to add a Slack interface or log conversations to a Google Sheet, simply add a brand new node to your workflow.
When you’re beginning your n8n journey and wish inspiration, be happy to discover my templates.
About Me
Let’s join on Linkedin and Twitter; I’m a Provide Chain Engineer utilizing knowledge analytics to enhance Logistics operations and cut back prices.
For consulting or recommendation on analytics and sustainable Provide Chain transformation, be happy to contact me through Logigreen Consulting.
Samir Saci | Knowledge Science & Productiveness
A technical weblog specializing in Knowledge Science, Private Productiveness, Automation, Operations Analysis and Sustainable…samirsaci.com
