With gold prices reaching unprecedented levels in 2024, jewellers have faced increasing pressure on margins and inventory management.
Recent market volatility has highlighted the critical importance of accurate forecasting for businesses managing precious metal stocks, particularly given the sector’s traditional requirement for advance production planning and seasonal buying.
Understanding likely market movements has become more crucial than ever, as businesses seek to balance stock holdings against working capital while maintaining competitive pricing.
Against this backdrop, new analysis from BullionVault has revealed that artificial intelligence tools may offer jewellers more accurate precious metal price predictions than traditional forecasting methods.
The study shows that Google’s Gemini artificial intelligence tool demonstrated superior forecasting accuracy throughout 2024 compared to both ChatGPT and traditional human forecasters. This development could provide jewellers with an additional tool for planning purchasing decisions and managing risk in an increasingly volatile market.
Methodology and Scope
The research compared quarterly price predictions from two leading AI platforms against forecasts from BullionVault’s investor community of 2,002 respondents and London Bullion Market Association (LBMA) analysts.
The study encompassed gold, silver, platinum, and palladium prices across all four quarters of 2024, providing a comprehensive view of AI forecasting capabilities in the precious metals market.
Gold Market Performance
Google’s Gemini demonstrated exceptional accuracy in gold price predictions, with forecasts varying only 1.7% from actual quarterly prices throughout the year. This accuracy improved to a 0.4% variance in the final three quarters. The AI tool predicted ranges of $2,120-$2,344 for Q1, rising to $2,500-$2,800 for Q4, closely matching the actual quarterly averages of $2,071.76 to $2,661.61.
For manufacturers and retailers, these price increases created significant challenges in maintaining margins while meeting key retail price points. Businesses using forward buying strategies based on these predictions could have better protected their margins compared to those buying at spot prices.
Silver Market Insights and Stock Management
ChatGPT showed particular strength in silver predictions, with an 8% average variance from actual quarterly prices. The platform’s more conservative approach, forecasting ranges of $20-$30 in Q1 to $21-$33 in Q4, proved more accurate than Gemini’s bullish predictions of $25-$30 rising to $35-$40. The actual prices averaged $23.36 in Q1, reaching $31.36 in Q4.
These price movements particularly affected fashion jewellery manufacturers, who typically work with larger silver quantities and tighter margins. The accuracy of these predictions could help businesses better plan seasonal production schedules and manage stock levels across different product categories.
Platinum and Palladium: Manufacturing Considerations
Gemini’s platinum forecasts maintained a 1.0% variance from actual prices, demonstrating consistent accuracy across different metals. The tool’s more conservative outlook compared to ChatGPT’s wider ranges proved more reliable, with actual prices averaging between $909.63 and $981.20 throughout the year.
The palladium market highlighted the potential limitations of AI forecasting, with ChatGPT predicting prices more than double the actual market rates. This demonstrates the importance of using multiple data sources when planning purchases of specialist metals, particularly for manufacturers working with specific alloy requirements.
Practical Applications for Jewellers
These findings suggest several key strategies for jewellers to implement in their business planning. In terms of stock management, businesses can use these quarterly predictions to make informed adjustments to their inventory levels throughout the year. Seasonal buying can be planned to align with predicted price movements, while stockholding can be balanced across different metals based on the demonstrated accuracy of predictions for each metal type. Jewellers may also benefit from developing specific strategies for different metals, taking into account the varying reliability of AI predictions for each.
Risk management strategies can be enhanced by using these predictions to guide forward buying decisions and inform hedging strategies for larger stock commitments. Different approaches may be needed for various metals, with particular attention paid to building contingencies for less predictable metals such as palladium.
Price setting can be approached more strategically, with businesses developing flexible pricing strategies that account for predicted market movements. This includes planning price adjustments around forecast changes and implementing metal-specific markup strategies based on price volatility. Long-term orders may require specific contingencies for price protection.
2025 Market Outlook and Planning
BullionVault’s latest survey presents an optimistic outlook for 2025, forecasting gold at $3,070 per ounce, silver at $36.80 per ounce, platinum at $1,037 per ounce, and palladium at $1,076 per ounce. These predictions suggest continued upward pressure on costs, making it essential for jewellers to review their pricing structures to maintain margins. Businesses should consider adjusting stock levels accordingly, evaluating their metal combinations in products, and planning manufacturing schedules around predicted price points. The emergence of AI as a reliable forecasting tool marks a potential shift in how the jewellery trade approaches market analysis and planning. While AI tools have demonstrated impressive accuracy, particularly in major metals like gold and platinum, their integration into business strategy should complement rather than replace traditional market analysis and human expertise.
The study suggests that jewellers might benefit from incorporating AI forecasting tools into their business planning, while maintaining awareness of the technology’s current limitations, particularly in less liquid markets.
