Companies are accelerating remanufacturing as a way to mitigate supply chain shortages, reach new customers through affordability, and implement high-margin alternatives for parts.

However, those looking to build or optimise their remanufacturing operations face unique challenges, such as pricing a long tail of SKUs and undertaking accurate core forecasting—that is, predicting the volume, timing, and quality of returned products (core) that will be available for remanufacturing.

Enter AI.

As the costs of cloud storage and processing and prepackaged tools decrease, AI is becoming more accessible for organizations, helping them improve efficiency, yield, and margins.

Unlocking AI in remanufacturing

AI is an umbrella term for leading-edge techniques and methods of analysing data, predicting outcomes, and generating insights. While AI and gen AI technologies have garnered a lot of attention, they are only part of the AI landscape.

Core forecasting

The unpredictable nature of core availability is a challenge for remanufacturers, and conventional analytics lack the sophistication needed for predictive analysis.

Introducing a specialised system such as a forecasting tool could help remanufacturers evaluate core availability at a per-SKU level.

This forecasting tool could be trained on historical part performance data to assess the following:

• the estimated useful lifetime of a part
• the historical use rate of a part (such as hours or miles per day)
• macroeconomic conditions by region and industry that may affect trade-in timelines

Adopting AI tools for core forecasting could help remanufacturers reduce core safety stock by 2 to 4 percent and save 3 to 5 percent in freight costs by reducing the cost of expedited shipping.

AI tools could also help remanufacturers reduce overtime expenses, lose fewer sales due to stockouts, and ensure they have the parts that end customers need.

Case study: Incentives for core availability. A top technology OEM struggled to match the regional availability of core to demand. The refurbishing team harnessed an AI ecosystem that included a forecasting layer, sourcing algorithm, and valuation algorithm.

At the customer level, the team was able to identify lifetime value and historical purchases by end customer. At the product level, they tracked battery health, service history, and usage rate. They then used all this data to create a personalized trade-in offer based on each customer to help ensure that core was available when the OEM needed it (exhibit).

Pricing optimisation

Pricing analytics are inherently intricate and complex because of the high number and long tail of SKUs, tier product dynamics such as OEM-grade or tier-two products, status codes, subcore, and machine vintages. AI gives organizations multiple opportunities to optimize their pricing, including the following:

• optimising portfolio prices through AI (such as microsegmentation and cross-SKU demand optimization)
• reducing cannibalisation
• achieving more granular pricing (for example, per SKU, per customer segment, and per vintage and condition)
• adopting smart data fill (such as combining similar SKUs to compensate for data gaps)

Case study: Using machine learning tools to price long-tail SKUs. 

One independent remanufacturer’s portfolio includes a high proportion of long-tail, low-velocity, niche SKUs that transact less than once per year. It typically used a broad, rule-based approach for pricing such SKUs.

The remanufacturer harnessed AI and machine learning tools to identify the top factors influencing pricing power, such as manufacturer brand, shipping requirements, and customer characteristics.

It deployed analytical AI to determine the optimal prices for individual SKUs and reevaluated price changes within various scenarios to fine-tune the pricing.

By implementing these approaches, the organisation increased its profit margins by 11 to 15 percent, and an automated AI tool allowed it to price more than 140 million parts in real time.

AI-enabled warranty claims management

Managing warranty claims is often a considerable challenge for remanufacturing organisations, given the high volumes of unstructured text data, and it is exacerbated by the high volume of SKUs.

Large language models (LLMs) can help identify and evaluate recurring patterns in text data for warranty claims. Powered by gen AI, LLMs find the most frequently occurring contextual word combinations, produce insights based on those results, and generate reports that can be used by the warranty claim teams and R&D teams alike.

Gen AI can find patterns to reduce warranty costs and help companies better engineer components for long-term performance by doing the following:

• extracting a detailed report of breakdown events
• enabling rapid intervention for process and R&D faults
• identifying concurrent conditions among service bulletins
• identifying long-term part failure points
• generating specialized reports for product and return material authorisation teams
• cost-effectively analysing the long tail of claims

Implementing gen AI in warranty claims could yield a 5 to 10 percent decrease in warranty costs, according to McKinsey analysis.

Case study: Using remanufacturing warranty claims to improve long-tail part performance and capture market share.

A global OEM seeking to increase its share in the US market needed to differentiate itself from incumbents. To do so, the OEM used remanufacturing warranty data to analyse the long-term-wear patterns of remanufactured parts.

The OEM then fed the information from its analysis back into R&D and improved the long-term reliability of its new and remanufactured components.

The resulting overall increase in reliability benefited both the end consumers and the OEM. End consumers were able to increase their uptime hours by about 7 percent because of decreased maintenance downtime, the OEM’s warranty provisioning costs dropped by about 25 percent, and the OEM was able to command a higher price premium for its products.

A path to stronger AI capabilities

Algorithms and data are only part of what is needed to drive real value with AI. Effectively strengthening the analytics muscle often requires building capabilities beyond technology.

The remanufacturing teams that invest as much time and resources in the organisation and change management needed to support the technology as they do in the technology itself are best positioned to fully reap the benefits of AI.

Remanufacturers can create a solid foundation for AI by taking the following steps:

1. Focus use-case implementation.
2. Secure a business sponsor.
3. Align priorities with top management.
4. Assess organisational capabilities and needs.
5. Create a well-thought-out road map for implementation.
6. Double down on change management.