In the earlier days, vehicle analysis was a predominantly manual process. Mechanics would rely on their experience to identify issues in the vehicles. They would check every car component manually to detect the problem. Besides being a reactive approach, it was also time-consuming. The R&D and quality teams would often lack the data needed to improve the vehicle’s efficiency or innovate.
However, cut to the present, vehicle maintenance and research have become more efficient and proactive. The modern vehicles are no longer just modes of transportation. They are now the carriers of data that could help automobile manufacturers innovate and improve the safety and efficiency of vehicles.
All of this has been made possible by connected vehicle data. Modern vehicles provide real-time data about the vehicle’s condition, making it easy for the automobile OEMs to identify issues before they occur and offer personalized value-added services.
However, connected vehicles also have their own challenges. One of them is data retrieval and processing.
In this blog, we’ll explore it and discuss how Ascentt helped a global automobile OEM save costs and time with the Connected Vehicle Data Optimization Agent.
Use Case – Connected Vehicle Data Optimization Agent Cuts Data Processing Time
https://youtu.be/5Mp3ETgAU8M?si=F7awZxmCDrUnQqQ-
Problem: Long Processing Time And Escalating Costs
One of the leading global automotive OEMs wanted to provide its R&D and quality engineers with real-time vehicle data to improve quality and accelerate innovation.
However, it wasn’t simple. The automotive giant faced the following challenges in retrieving and processing data:
- Data volume: The OEM had to retrieve data from various sources, including sensors, cameras, and communication systems. This meant they were dealing with an enormous amount of data from thousands of connected vehicles. This made it difficult for them to store, retrieve, and process the data. The sheer volume of data slowed down the turnaround time for meeting thousands of critical data processing requests.
- Data processing costs: The computing cost of processing a single data request ran into millions of dollars. Every data processing request involved allocating resources to access storage and network layers to retrieve data. A significant portion of the budget was allocated to operations and infrastructure maintenance rather than innovation.
- Data processing time: It took over 4-8 hours, sometimes even four days, for the OEM to process each data request. This prevented the R&D and quality teams from receiving real-time data, which could have helped them resolve critical product issues or make design iterations.
- Scalability limitations: The OEM had to process plenty of data requests for thousands of vehicles. With each request taking 4-8 hours, meeting the personalized needs of each vehicle became challenging.
The Solution: Ascentt’s Vehicle Data Optimization Agent
Ascentt built a Vertical AI solution to cut computing costs and reduce processing time.
What we did:
- Reviewed the existing data storage architecture and the data retrieval process to understand the root cause of the challenges.
- Determined that the existing systems were not designed to process the large volume of data generated by modern connected vehicles.
- Developed a customized data retrieval algorithm to extract relevant insights required for analysis, reduce computational burden, minimize cloud resource consumption, and provide real-time data to engineering and R&D teams.
- Deployed an unnesting technique to extract actionable insights from connected vehicles and ensure only relevant data is retrieved quickly.
- Built different data retrieval algorithms for different compute sizes and workloads.
- Integrated connected vehicle data with OEM’s enterprise system to enhance customer experience and quality.
Key capabilities:
- Data unnesting technique: Extracted actionable insights from nested objects to accelerate data processing.
- Dynamic compute sizing: Chose the right compute size based on the workload.
Result: From 96 Hours To Under 4 Hours
The impact of our solution was almost immediate.
- The OEM could save millions of dollars in computing data requests, as there was no over-provisioning of resources to retrieve data.
- The data retrieval time was reduced from 96 hours, or four days, to just four hours.
- The R&D and quality engineering teams could focus on making data-driven decisions to boost quality and customer experience, rather than retrieving data.
Why Vertical AI?
In a high-throughput scenario like the automotive industry, OEMs require a more sophisticated AI solution than off-the-shelf AI.
They need a custom-built AI that is purpose-driven and specially developed for the automotive industry’s shop floors, systems, and environments.
That’s why we at Ascentt build vertical AI solutions.
The data optimization agent is a part of our vertical AI solutions. We built it to help the automotive giant enhance quality and accelerate innovation without spending millions of dollars or time.
Unlike generic AI, vertical AI models are pre-trained on the company’s existing manufacturing workflows. This ensures that the solutions are designed based on the company’s business needs.
Planning To Save Millions of Dollars and Hours On Data Processing?
Ascentt’s Data Optimization Agent is already showing measurable results in managing large volume data requests for connected vehicles. It is helping automobile OEMS:
- Reduce the turnaround time for processing data requests
- Optimize cloud budget and reduce computing costs
- Transform customer experience with real-time data
Want to see it in action? Talk to our AI experts today.
Your next data processing could happen in four hours instead of four days.
