Connected Worker Exchange: Hao Dinh AI Use Cases
May 5
5 min read
0
83
0
At the Connected Worker Conference, we explored real-world AI use cases transforming manufacturing, operations, and supply chain. These weren’t just concepts—they were actionable examples with measurable impact. Below, you’ll find step-by-step instructions to prototype each use case. By building these prototypes, you’ll uncover the real value AI can deliver, understand the types of data required, and gain hands-on experience developing AI solutions. Whether you're optimizing workflows, predicting maintenance, or automating decisions, prototyping helps bridge the gap between AI theory and business results. Dive in and see how AI can drive productivity, efficiency, and smarter decision-making in your organization.
Automate Data Collection using AI:
Business Value: AI automates data entry by extracting and digitizing information from paper forms, machines, and sensors—reducing manual work, errors, and delays. This boosts speed, accuracy, and productivity in factories. By eliminating repetitive tasks, workers can focus on higher-value activities, driving smarter decisions and more efficient operations. (Sources: McKinsey BCG PwC)
Let's have some fun using AI to scan documents to automatically digitize data collection.
1: Click on this link: Google Document AI
2: Scroll down to the "Try Document AI in your environment" section (see below picture area highlighted in red).
3: Click on the "Invoice" sample document.
4: Click the "I'm not a robot" checkbox.
5: After a few seconds, navigate through the 'FIELDS', 'KEY VALUE PAIRS', and 'OCR' tabs to see how AI has scanned the paper invoice and digitally extracted key data points like 'Invoice Number', 'Total Price', and 'Ship to Address.'"
Imagine using AI to instantly scan and digitize any document—eliminating the need for manual data entry and reducing errors. Once digitized, the data can seamlessly flow into other systems, enabling automated actions without ever having to re-enter information
Decrease Defects using AI:
AI is revolutionizing quality control in manufacturing. By training models to detect defects and deploying vision systems for real-time inspection, companies can reduce defects by 25% to 50%. This not only improves product quality but also cuts waste, boosts efficiency, and enhances customer satisfaction across the production line. (Sources: McKinsey BCG PwC)
Learn how to turn your webcam into a banana ripeness detector with Google’s Teachable Machine! This fun tutorial walks you through training an AI model to classify bananas as unripe, ripe, or overripe. No coding needed—just upload images, train your model, and test it live. A perfect hands-on intro to using AI to detect defects.
Click Here for Tutorial: Bananameter | by Barron Webster
The banana AI tool offers a simple way to understand how vision AI can be trained to detect visual differences. By applying the same approach, manufacturers can create custom vision AI solutions to automatically inspect products for defects like scratches, dents, or missing components - reducing errors, increasing efficiency, and ensuring consistent quality control.
Decrease Defects using AI:
AI is transforming factory safety by predicting incidents before they happen. By analyzing sensor data, operator behavior, and maintenance logs, AI identifies risks like equipment failure or PPE non-compliance. This enables real-time alerts and preventive actions—reducing injuries, improving compliance, and fostering a safer, smarter manufacturing environment. (Sources: BCG PwC
Follow the below instructions to experience how AI can help reduce safety incidents.
Step 1: Download sample safety incident data
🗂️ Sheet 1 Details: Incident Logs
Field Name | Description |
Date | Date of the incident occurrence |
Shift | Shift during which the incident occurred (A/B/C) |
Machine ID | Identifier for the machine involved |
Operator ID | Identifier for the operator involved |
Incident Type | Nature of the incident (e.g., Slip/Fall, Burn) |
Severity | Severity level (Low/Medium/High) |
Days Since Last Maintenance | Days passed since last machine maintenance |
Weather | Weather condition at the time (could influence slips, etc.) |
📊 Sheet 2 Details: Sensor Data
Field Name | Description |
Timestamp | Date and time of data capture |
Machine ID | Machine being monitored |
Temp (°C) | Machine surface or internal temperature |
Vibration (Hz) | Vibration frequency—can indicate wear/malfunction |
Noise (dB) | Decibel level—high noise may suggest mechanical problems |
Humidity (%) | Ambient humidity level, which can impact safety and equipment performance |
Operator Present | Whether a human operator was present during data capture |
👷 Sheet 3 Details: Training Records
Field Name | Description |
Operator ID | Identifier for the operator |
Last Safety Training | Date the operator last completed safety training |
PPE Compliance Rate (%) | Percentage of time the operator used personal protective equipment correctly |
Incident History | Summary of past incidents involving this operator |
Days Since Last Training | Number of days since last safety training session |
Step 2: Ask AI Questions
If you have access to Microsoft Copilot or ChatGPT, upload the sample data and try asking: "From the attached data, what are the top safety issues"
Let's use Microsoft Copilot. Click here for instructions to access Copilot.
In Copilot, click the "+" icon and select "Upload" to upload the sample file
Find the sample file and click "Open"
Type or "cut & paste" into the Message Copilot section: From the attached data, what are the top safety issues; click the "up arrow" button
AI reviews the sample data and generates the following insights. Note, your results could vary due to the AI used.
AI is not infallible—always review and validate its recommendations. If the insights are viable, take appropriate action. If the output isn’t useful, consider refining your prompt (question) or providing better data to improve the quality of the AI’s response.
Optimizing Inventory using AI:
AI helps optimize inventory by accurately forecasting demand, dynamically adjusting reorder points, and minimizing stockouts or overstock. With real-time data analysis, AI reduces holding costs, improves service levels, and ensures the right products are available at the right time—leading to smarter, leaner, and more responsive supply chain operations. (Sources: McKinsey BCG PwC)
Let’s bring AI into your inventory decision-making!
Step 1: Set the Scenario
Imagine you manage inventory for an online sneaker store. Your best-selling model, the “AI Runner 3000,” is flying off the shelves, and you need to decide how many to reorder for next month.
Step 2: Gather Your Data
Let's assume:
Last 3 months’ sales: 500, 700, and 1,000 units of AI Runner 3000
Supplier lead time: 2 weeks
Average seasonal demand increase: 20% in summer
Step 3: Use Generative AI to Predict Demand
If you have access to Microsoft Copilot or ChatGPT, try asking:
"Based on past sales of 500, 700, and 1,000 sneaker units, a supplier lead time of 2 weeks, and a 20% seasonal demand increase, how many sneakers should I order for next month?"
The AI will likely suggest a number based on trend forecasting—similar to how real companies optimize their supply chains!
Let's use Microsoft Copilot. Click here for instructions to access Copilot.
In Copilot, ask the following question
Type or "cut & paste" into the Message Copilot section: Based on past sales of 500, 700, and 1,000 sneaker units, a supplier lead time of 2 weeks, and a 20% seasonal demand increase, how many sneakers should I order for next month?
Hit "Enter" or click the "Up Arrow" (See below):
Copilot will do its best to answer your question. However, responses may vary as Copilot continuously learns and adapts over time. Below is a sample reply.
Step 4: Compare Generative AI’s Answer vs. Your Gut Feeling
Was the AI's prediction higher or lower than your initial guess?
What other factors (marketing promotions, competitor activity) should you consider?
Could AI help automate this process for you in real life?
🎯 Final Thoughts: AI is Here to Help, Not Replace
Generative AI isn’t just a buzzword—it’s a powerful tool that helps businesses make faster, smarter decisions. It won’t replace inventory managers, but it will augment their ability to forecast, optimize, and automate operations. 🚀
