Xometry Upgrades AI Model Architecture Across Quoting and Sourcing

Smarter process recommendations from the first upload

Xometry’s upgraded context-aware AI process recommender understands a part’s industry application – recognizing, for example, that an aerospace component likely calls for tighter tolerances and flight-grade alloys – along with its probable material and geometry, closing the gap for first-time customers who have no order history for the model to draw on. Buyers receive a manufacturing recommendation when they upload a part, providing them with the optimal process from among 20 supported manufacturing techniques. Early feedback has been positive, with the AI recommendation being accepted by buyers more than 85% of the time – a marked improvement over the prior model, with the biggest gains among first-time customers.

A new generation of cost models

The way Xometry prices custom jobs is evolving. The new generation of cost-prediction models prices each part based on the specific parameters required to manufacture it, considering a greater breadth and depth of inputs, including geometry, material, finish, and whether the part is a standalone or part of a larger job. This granular input leads to more accurate pricing.

Starting with CNC machining, the new cost-prediction models are trained on Xometry’s proprietary data insights and understand the specific parameters each part requires, delivering an approximately 15% improvement in CNC cost-prediction accuracy, validated through live testing. The models predict a calibrated cost distribution for every CNC job, so Xometry can gauge and signal its confidence level before a quote is returned.

The company plans to extend the upgraded models to additional processes within Xometry’s Instant Quoting Engine, and to incorporate additional CNC materials directly into the engine. As new materials are added, the models will continue to improve, built on a broader set of historical custom material requests.

Faster sourcing, better-fitting jobs for supplier partners

The new adaptive sourcing models leverage Xometry’s proprietary supplier data layer to price each job dynamically, moving quickly on well-understood jobs. The models also incorporate an upgraded job-partner suitability that scores every job against a supplier partner’s machine characteristics – including dimensions of the machine and sub-process capabilities – in addition to quality history and on-time shipping record. This has significantly improved partner-matching. Rather than navigating a noisy board of broad job notifications, partners are presented with a curated flow of better-fitting opportunities. This combination of deep capability matching and responsive pricing leads to a stronger network overall.

For more information, visit [xometry.com](https://www.xometry.com/).