MAIN CHALLENGE: Zero-Fail Logistics (Valio Aimo)

SIDE CHALLENGES: Best Use of Google Cloud. Best Use of Gemini API.

Valio AI-mo (End-to-End, AI-driven logistics platform) designed for food supplier companies.

Our solution consist of 3 main components integrated to a unified workflow:

1. Our Predictive analysis machine learning model trained using Google Cloud Platform. We trained 3 different machine learning algorithms and decided to use random forest (most accurate predictions) at the end. The model uses order, stock, and supplier/quality data to forecast which items are likely to be missing.

2. Organization side platform allows for resource management with AI-assisted operational decision making and automation. The platform is done with vue.js. The platform integrates the predictions from our model and the customer input (eg. missing order complaints).

3. Customer side is handled by a conversational, multilingual, decision-making AI agent powered by Gemini. Customer communication is handled by Telegram in our prototype. It uses NLP/NLU to deliver immediate responses to customer complaints.

Team name: Hungry&Finished

Bátor Lang, Eeli Remes, Márton Magyar, Aarni Viljanen, Eeli Ogata

In practice, Valio AI-mo works by streamlining logistics for food supplier companies through a fully integrated, end-to-end solution that combines predictive analytics, operational automation, and customer interaction in a seamless workflow. Here's how each component operates in real-time:

Predictive Analysis (Machine Learning Model)

• Data Inputs: The machine learning model (random forest) is trained with data such as order patterns, stock levels, supplier performance, and quality data.

• Prediction Process: The machine learning model can predict future data (like incoming orders, stock quantities, and order spikes).

• Outcome: The machine learning model predictions can be used to react to future supply chain issues. In the future the system could automatically order foods in advance if it predicts that there could be a shortage.

Organization Side Platform (Operational Decision-Making & Automation)

• AI-Assisted Resource Management: When the system predicts a potential shortage, the platform could recommend specific actions (like prioritizing orders for customers who are more likely to experience disruptions).

• Workflow Integration: Customer complaints are automated into the workflow system. AI can be used to get a summary of the overall situation of missing orders.

• Automation: Tasks like reordering supplies, sending notifications to suppliers, or updating inventory levels can be automated, freeing up human operators to focus on higher-level decisions.

• Decision Support: The platform shows the predictions and real time situations to help with decision making. AI will also help with decision making.

Customer Side (Conversational AI Agent)

• Conversational AI: Customer will get automated answers from AI-powered conversational agent powered by Google's Gemini and NLP/NLU technologies. The AI agent can handle talking with multiple languages to give the customer the best customer service experience.

• Customer Interaction: The AI agent analyzes the complaint and cross-references it from the machine learning model to offer a solution. AI agent can have limited decision making capabilities like making orders for customers.

• Real-Time Updates: The AI agent can give customer the most up to date information and provide quick + realistic responses, which improve customer satisfaction.

• Learning and Improvement: The AI Agents learns from customer interactions and it improves by time.