其他行业 · AI应用
预测性维护试点项目
Motability Operations（英国一家为超过70万残疾客户提供服务的机构）因意外故障面临车辆可靠性和客户满意度方面的挑战。为此，他们启动了为期六个月的预测性维护试点，利用AWS服务（特别是Amazon SageMaker）分析联网车辆数据并实施AI/ML解决方案。这一创新方法成功预测了49%的潜在故障，并纠正了七个主要车辆故障，显著提升了运营效率。试点期间客户信任度保持高位，退出率低于0.5%，最终改善了整体客户体验。
故障预测率
49%
来源披露
纠正的重大车辆故障数
7
来源披露
客户退出率
低于0.5%
来源披露
01
业务背景
Motability Operations（英国一家为超过70万残疾客户提供服务的机构）因意外故障面临车辆可靠性和客户满意度方面的挑战。为此，他们启动了为期六个月的预测性维护试点，利用AWS服务（特别是Amazon SageMaker）分析联网车辆数据并实施AI/ML解决方案。这一创新方法成功预测了49%的潜在故障，并纠正了七个主要车辆故障，显著提升了运营效率。试点期间客户信任度保持高位，退出率低于0.5%，最终改善了整体客户体验。
02
遇到的问题
Unexpected vehicle breakdown is always inconvenient, but for Motability Operations’ customers, it can be significantly more challenging. Motability Operations wanted to reduce the number of breakdown incidents by using connected vehicle data.
03
AI 解决方案
Motability Operations collaborated with AWS Professional Services to deliver a 6-month predictive maintenance pilot. They collected connected vehicle data over six months to train foundational ML models, then used Amazon SageMaker to build models predicting whether a vehicle was likely to break down soon. The solution employed serverless and managed AWS services to achieve flexibility, scalability, and cost-efficiency.
- 1与AWS Professional Services合作建立初始概念验证
- 2选择并使用多种AWS无服务器和托管服务以搭建架构
- 3花费6个月时间收集并训练数据，构建基础机器学习模型
- 4利用Amazon SageMaker构建和部署模型，判断车辆是否可能在近期发生故障
- 5在6,500辆车上进行试点，预测故障并主动干预
04
实施结果
故障预测率：49%（未披露）
纠正的重大车辆故障数：7
客户退出率：低于0.5%（低退出率保持客户信任）
05
风险与边界
未披露
“The trial was more successful than we had imagined. We were able to intervene proactively, manage cases, and enhance the customer experience.”
Felicity Kelly · Connected and Digital Innovation Lead
信息来源
1 个来源页面内容为结构化改写；关键结论应可追溯到以下材料。