Transforming fleet management with AI using Amazon Quick with Astrata

Astrata’s transport management solutions provide customers with data, such as driver performance, maintenance schedules, and vehicle diagnostics, to monitor and manage fleet vehicles. But operations managers still had to manually review data and analyze the findings to gain actionable insights. Considerable time was spent compiling monthly and quarterly reports, while accident investigations took up to a month. This increased cognitive workload, leaving less time to improve fleet safety and efficiency. Drivers could also spend lengthy periods on administrative leave while accident investigations were completed.

Astrata saw an opportunity to make operational data easier to navigate by using AI. The company wanted to collect relevant information and present it through clear dashboards and tailored reports so that users could absorb information easily and make decisions faster. Astrata also wanted a simple-to-use chat function so that operations managers without specialist knowledge could find answers in the data more easily and avoid issues, such as maintenance failures and inefficient fleet performance.

Astrata worked alongside AWS, with support from the Singapore government’s Enterprise Compute Initiative program, to build Astrata Intelligence quickly. The company developed an initial test site in less than a month and launched the customer-facing version in a similar time frame.

“Our focus has always been putting our customers front and center. Using AWS means we can focus on the applications of our solution for our customer while knowing that the team at AWS will focus on the foundational technology that we can build upon,” says Marius van den Berg, chief technology officer at Astrata.

Astrata Intelligence helps users unlock the value of fleet data by making relevant information easier to find and understand. Its dashboards draw from a single source of information and present data clearly so that users can assess fleet operations and make decisions faster. Automated reports are tailored to specific requests and audiences, combining all relevant details in a single response. A chat function also gives nontechnical employees a way to query the data in plain language and receive immediate insights and recommendations.

Astrata Intelligence’s natural language querying, near real-time dashboards, and automated report generation are supported by Amazon Quick, which answers questions and turns those answers into actions using agentic teammates. Customer data is merged with broader industry data using Amazon Bedrock, the solution for building generative AI applications at scale, to provide industry benchmarking and tailored recommendations.

Astrata hosts Astrata Intelligence on Amazon Elastic Compute Cloud (Amazon EC2), which offers secure, resizable compute capacity in the cloud. Amazon CloudWatch, an intelligent observability service that provides actionable insights across applications and infrastructure, supports performance and security. And the processed data is stored on Amazon Relational Database Service (Amazon RDS), an easy-to-manage relational database service optimized for total cost of ownership.

Astrata Intelligence has cut accident investigation time from around a month to just 1–2 days. This isn’t just an efficiency gain; it’s a human one, too, with drivers needing to spend much less time on administrative leave while those investigations take place.

In addition, customers can now shift from reactive to proactive maintenance. Astrata Intelligence has a 91 percent prediction accuracy for maintenance failures, and one customer noted 14 vehicles being flagged as likely to fail within the following 30 days. Three incidents were avoided due to predictive alerts.

Operations managers also spend less time compiling monthly and quarterly reports and can retrieve insights that combine multiple factors through a single query. This reduces their cognitive workload, giving them more time to focus on fleet safety and efficiency.

“With Astrata Intelligence on AWS, we have a very well-trained, well-informed, highly skilled entity that you can just go and ask, and it will come back with an answer,” says Van den Berg.