Sunday, June 16, 2024

Striim and BigQuery Power’s UPS AI-Driven Delivery Security


UPS Capital is using artificial intelligence (AI) and Google’s Data Cloud to protect parcels against porch piracy. Over 300 million American consumers now purchase online, and UPS Capital has seen that protecting their package delivery ecosystem presents serious issues for their clients. The organisation is currently using data access and digital capabilities to assist clients reconsider conventional methods of preventing shipment loss and improving customer experiences.

Using AI and real-time data to secure shipments

Originally founded as a financial services branch in the 1990s, UPS Capital has significantly evolved into a comprehensive suite of solutions aimed at safeguarding businesses’ assets while they are in transit. With a variety of products and services, such as cutting-edge insurance, enhanced carrier protection, and assistance with integration for leading e-commerce platforms, UPS Capital enables companies to manage shipment risk-free and profitably while maintaining cash flow.

Shipping loss and insurance claims have increased due to the exponential rise in package delivery and the higher value of requested commodities, which creates an appealing opportunity for criminals. According to a TransUnion research from 2022, shipping fraud increased by more than 1,500% just from 2019 and 2021. UPS Capital has long provided clients with insurance solutions that act as a monetary safety net against shipment loss. UPS Capital has been able to adopt a new strategy for risk reduction with the predictive intelligence of DeliveryDefense Address Confidence because to the development of computing power and artificial intelligence.

Using machine learning algorithms and real-time data, DeliveryDefense Address Confidence protects deliveries by utilising Google’s Data Cloud and AI technology. It improves the evaluation of successful delivery probability while reducing the danger of loss or theft by giving probable delivery sites a confidence level. On a confidence score range of 100 to 1000, where 1000 denotes the highest likelihood of a successful delivery, each address is assigned a number. Based on customer reports of parcel theft, these scores are calculated. Through an API, shippers can include this score into their shipping workflow to take preventative, proactive measures on addresses with low confidence.

For example, the merchant has the option to proactively reroute the cargo to a secure UPS Access Point site if the package is intended for an address with a low confidence score. Because of their stringent chain of custody security protocols, these facilities usually earn a confidence score of approximately 950.

“BigQuery, Looker, and Vertex AI are being used by UPS to develop products that shield clients from danger by assisting them in preventing it in the first place. Customers can save money through lower losses, save time by reducing the number of human order reviews, and enhance their overall shopping experience by using DeliveryDefense Address Confidence scores in the shipping process. – Pinaki Mitra, UPS Capital’s vice president of data science and machine learning.

With DeliveryDefense, SunJoy, a customer of UPS Capital, has decreased its losses by 35% by adding an adult signature or rerouting items to a safer area in response to notifications regarding high-risk deliveries.

This service not only promotes customer happiness through timely and dependable deliveries, reduced costs, and overall better service quality, but it also streamlines operations and improves delivery fulfilment rates. This has a knock-on effect that affects every party participating in the supply chain favourably and makes the market more resilient and competitive.

The core technology of UPS DeliveryDefense

UPS Capital Delivery Defense and Claims Investigation
Image credit to Google Cloud

Businesses can more easily manage their data with the help of Google Cloud’s unified data cloud, which streamlines the entire data lifecycle. Organisations may gain important insights without needless complication by employing cutting-edge technologies like Vertex AI and integrating various datasets into BigQuery with ease.” -John Kutay, Striim‘s Director of Product, Data, and AI.

DeliveryDefense Address for UPS A strong technical foundation is necessary for confidence, and it starts with the direct input of various datasets into BigQuery. BigQuery is Google Cloud’s cornerstone structured data repository, offering a dependable data warehousing solution. Concurrently, information obtained from SQL Server is carefully cleaned. These harmonised datasets are pooled in BigQuery and are easily connected with other Google Cloud services. They have been enhanced with multimedia components.

Subsequently, Vertex AI emerges to enable the implementation of complex machine learning models, including shipping transaction fraud detection and route anomaly detection models. With the wide range of tools provided by Vertex AI, UPS Capital can develop, hone, and apply models to uncover patterns, forecast trends, and extract insightful information from its data. Through the Looker API, Firestore, a flexible and scalable NoSQL database for server, browser, and mobile development from Firebase and Google Cloud, provides insights, confidence scores, and analytical details.

Real-time data streams utilising Striim and Google Cloud for generative AI

“With the help of Striim and Google Cloud, Google cloud can now use AI and ML to improve the customer experience.” – Pinaki Mitra, UPS Capital’s vice president of data science and machine learning.

At Striim, AI is driving innovation, especially in the fields of streaming contexts and gen AI. The real-time data integration platform enhances data representation, processing efficiency, and analytical correctness by dynamically embedding vectors into streaming information, complementing Google Cloud’s current design. Along with integrating data from many sources, both structured and unstructured, Striim uses a range of AI models from OpenAI and Vertex AI to create embeddings that produce similarity scores between data points to show potential correlations.

During this procedure, the streaming data is differentiated from conventional batch processing techniques by being enhanced with dynamically created vectors. In contrast to other approaches that depended on post-collection analysis, Striim‘s novel approach functions instantly, continuously enhancing and optimising the data flow. Through the integration of vectors into the streaming process, the system not only improves query relevance but also spurs continuous model improvement by bringing in real-time data.

Striim and Google Cloud were chosen by UPS Capital to make it possible to use AI models with real-time data streams. As a result, the organisation may lower the latency of identifying suspicious behaviour and continuously enhance its defence models at the speed of business. UPS Capital can improve the precision and efficacy of its defence mechanisms over time by analysing streaming data and dynamically creating vectors, guaranteeing continuous protection against shifting and emerging threats.

Future-Ready: UPS’s package delivery defence plan

Using real-time data and machine learning, UPS Capital’s DeliveryDefense Address Confidence, powered by Google BigQuery and Striim, revolutionises package delivery security. Shippers can reduce risk and enhance customer satisfaction by working with UPS Capital to implement proactive measures and give confidence scores. Learn how Striim on Google Cloud, like UPS’s DeliveryDefense Address Confidence, may enable real-time intelligence for AI. Get a free trial by registering now!

Since June 2023, Drakshi has been writing articles of Artificial Intelligence for govindhtech. She was a postgraduate in business administration. She was an enthusiast of Artificial Intelligence.


Please enter your comment!
Please enter your name here

Recent Posts

Popular Post Would you like to receive notifications on latest updates? No Yes