How AI parts prediction helped a global manufacturer to increase first time fix rates
Challenge
Our client, a leading home appliance OEM, manufactures and supports over 3,200 appliance models and ships around 28.2 million parts and accessories in the US annually.
The company had well established service management systems, tools and processes. However, no matter how efficient the agents and technicians were, if the right parts weren’t available, issues couldn’t be fixed the first time.
The high complexity and specialized nature of parts data, including parts BoMs, service records, and warranty claims, often residing in disparate systems made it challenging for the OEM to identify and provide the correct parts needed for each service visit, ultimately limiting the company’s ability to improve first-time fix rates (FTFR) and customer satisfaction.
Even with optimized service operations, the company had hit a ceiling on improving FTFRs as technicians continued to arrive on-site without the right parts. The core challenge wasn’t merely the availability of parts, but precisely identifying the correct parts to carry to fix the issue. Too often, this process relied on guesswork. These challenges negatively affected customer satisfaction (CSAT) scores and brand equity.
The business impact
- +
Stagnant FTFR: The company was stuck at an FTFR of between 75–80%, an inefficient rate for a leading global OEM.
- +
Multiple visits: Roughly 25–30% of service calls required multiple truck rolls due to lack of the right parts.
- +
Low productivity: Repeat visits were time consuming, affecting tech productivity and eroding profitability.
Solution implemented: Parts AI Agent
The solution was to deploy the Bruviti Parts AI Agent to combine various data sources across the OEM’s systems, then use the combined data to power specialized parts algorithms that accurately identify the parts needed to fix each issue. By making the issue the key that links all data, the models could pinpoint the right parts with high accuracy and confidence.
- +
Connecting complex disparate data: Using NLP augmented by LLMs, the Parts AI Agent integrated and harmonized diverse data sources like parts BoMs, service records, and warranty claims, transforming them into a unified, usable format to feed the predictive parts models.
- +
Specialized predictive algorithms: The Parts AI Agent used pre-trained neural networks to rapidly build predictive algorithms specifically tuned for parts, using extensive datasets unique to the customer. This enabled it to precisely identify the parts needed for each service issue right out of the box.
- +
Augmented scoring automation: Augmented scoring automatically optimized score performance, allowing the agent to adapt and learn from new data and make recommendations to continually enhance results for the company.
Solution deployment
The Parts AI Agent was embedded in the customer’s secure enterprise ecosystem and service applications, and deployed within existing call center and field operations. This plug-in flexibility ensured rapid deployment, with minimal training across the service organization.
Conclusion
The AI solution focused on the issue to accurately identify the necessary parts for repairs. By significantly improving parts identification accuracy, Bruviti transformed the OEM’s service operations, resulting in better-prepared technicians, increased FTFR, fewer truck rolls, and improved customer satisfaction.