Rogerio Pedral

Improved Manufacturing with Visual AI and Demand Forecasting

Lenovo is one of the largest technology companies in the world, invoicing more than $45 billion of computers, laptops, and accessories each year.

Rogerio Pedral leverages AI to help Lenovo innovate across their manufacturing processes, including quality control, predicting sell-out volume, and improved supply chain demand forecasting. This has improved their bottom line significantly, positioning them as a leader in their region.

One area on the forefront of data science is Visual AI, which Rogerio uses to automatically classify hardware, assisting humans working in their plant and shipping distribution centers. This has reduced errors and ensured that customers received the products they expected, thereby improving customer satisfaction and loyalty, as well as their company’s brand image. Rogerio’s use of Visual AI is an exciting and leading edge use case showing innovation and transformation for an established company.

Visual AI Classifications

Specific labels are placed manually and chosen by a person on the assembly line. Sometimes, people make a mistake and place an incorrect label, instead of an “Intel 5” sticker, they put “Intel 3”, for example. In the quality control (QC) process, a picture of the laptop and sticker are compared with the control picture to detect sticker application errors.

In the past, QC monitoring and decision-making were made by a person on an assembly line, which often resulted in an incorrect sticker or label, leading to returns, lower customer satisfaction, and even lawsuits. Now, with Visual AI technology from DataRobot, photographs of the laptop and sticker are compared to automatically detect sticker errors. DataRobot compares the photos taken for each laptop with those in the BOM image database and estimates the probability that the two images are the same. If the probability is less than a defined limit, there will be a red alert on the assembly line for the operator to review and correct. This process has improved accuracy from 93% to 98%.

Forecasting SellOuts

DataRobot has already helped Lenovo more accurately predict sellout volume — the number of units of a product that retailers sell to customers — Lenovo Brazil needed to build machine learning models at a faster rate and have those predictive models be more accurate. As a result, Lenovo has surged to become the leader in volume share on notebook sales for the B2C segment in Brazil.

For Lenovo Brazil, equalizing the supply and demand for laptops and computers among the Brazilian retailers that receive thousands of Lenovo products each week is a top priority. Having too much inventory means that stores will have to make special sales efforts to burn and move that stock. Having too little leads to the loss of potential sales due to a shortage of computers on the shelves.

Lenovo uses dozens of variables that could affect sellout volume at retailers – average product price, sellout rebate period, Google Analytics rankings, media campaigns, price differences between Lenovo and the competition. Previously, AI identified the specific variables that were truly predictive and had the most significant impact on estimating sellout volume, and then transparently communicated to business stakeholders the results of those models. Based on these findings, retailers working on huge promotional campaigns should be given a larger inventory on that given week, lest they be stuck with a shortage relative to customer demand.

The products are increasingly similar, the competitors more aggressive and competitive. We must guarantee the quality of our products and consequently the total satisfaction of our customers. We need increasingly intelligent tools to guarantee and improve quality, increase efficiency and reduce costs. In my opinion, AI comes to replace judgments, remove emotions, and maintain the quality standard from start to finish.