Using computer vision to broaden your vision

One of the biggest logistic companies teamed up with Valiance to build a computer vision service to detect people in the warehouse and create a more convenient way to make an inventory of goods.

The Challenge

The company stores a wide variety of products and has a lot of staff in their warehouses. They experienced frequent goods damage and goods loss every month. Thus, they have a security concern to better manage their warehouses.

The Solution:

We built a computer vision solution using image recognition model, face recognition model, and SKU classification model to automate warehouse surveillance and inventories management.

The Outcome

• Reduced goods damage and good loss by 5% per month.

If the media control the mind of the masses, then AI can help them to read the mind too.

Understanding the reader’s sentiment towards certain topics may help the media to produce better news. We worked with national mainstream media to create an information extraction and sentiment classification system using Natural Language Processing.

The Challenge

Media produce textual data (news) on a daily basis that potentially form public opinion towards certain topics. The ability to classify the sentiment towards news produced becomes a valuable asset for the company that helps them maintain the quality of the news and set a suitable tone.

The Solution

We developed an information extraction model and applied sentiment analysis with the latest technique called Natural Language Processing to automatically analyze various news stories and turn client’s textual data into digestible insights.

The Outcome

• Correctly classify 70% whether a certain news stories have neutral, positive or negative sentiment
• Cluster and summarize multiple news automatically.

Prevent bogus transaction, minimize the loss

Fraudulent transactions are common causes for cash leakage in retail businesses. In 2020, we worked with the biggest tobacco companies in Indonesia to prevent such issues from happening through AI.

The Challenge

A leading tobacco company in Indonesia wanted to mitigate monetary loss from B2B and B2C transactions in retail businesses.

The Solution

We developed automated anomaly detection and consistency monitoring to check every transaction.

The Outcome:

• Detect 90% of anomalous transactions
• Detect 25% of possible fraudulent transactions.

Leverage big data in designing a tourism strategy that works.

As Indonesia became one of the most popular destinations for tourism, We partnered up with a Big Data company to help one of government institutions to effectively bring more visitors to Indonesia.

The Challenge

The government institution did various marketing campaigns to maximize the number of foreign tourists in Indonesia. They also conducted analysis on historical data of tourist arrivals in approximately 150 gates, public holidays, and social media sentiments towards Indonesia every month. This process required so much time and effort if done manually.

The Solution

We developed an AI-powered tourism management system to make the data analysis process more efficient and to better optimize the marketing outcome.

The Outcome:

• Automate predictive analysis for number of tourists and marketing outcome with improved accuracy.
• Reduce data processing and data analysis time from days to 3 hours.