How Will AI Improve Manufacturing Recommendations?
The company said buyers accepted its automated manufacturing recommendations more than 85% of the time during early deployment.
Xometry, the AI-powered marketplace connecting buyers with manufacturing suppliers, announced on July 20, 2026 a broad upgrade to the artificial intelligence models powering its digital manufacturing marketplace. The company said the new architecture is designed to improve process recommendations, pricing accuracy and supplier matching.
The company said the enhancements are built on proprietary data spanning part geometry, manufacturability, pricing, supplier capabilities and production outcomes. This creates what the company described as a more connected intelligence layer across the manufacturing process.
Vaidy Raghavan, Xometry's chief technology officer, said the company's AI is constantly learning across the millions of parts quoted through the platform. "Our new models move beyond treating that knowledge as separate skills, where one model recommends how to build something, another prices it, another finds who should build it. Now, they are an intelligence layer that gets sharper every time a part moves through it," Raghavan said.
He added that the company's data models increasingly determine how parts are manufactured, what they cost and which suppliers are best suited to produce them. The company said this is intended to give customers more accurate recommendations, pricing and supplier matches with each new project submitted through the platform.
What the Models Do
According to Xometry, one of the most significant changes is an updated context-aware recommendation engine. The engine helps customers determine the most appropriate manufacturing process as soon as a part is uploaded, the company said.
Rather than relying primarily on a customer's order history, the model evaluates factors including a part's geometry, likely material selection and intended industry application, according to the company. The company said the system can recognize when aerospace components may require tighter tolerances or specialized alloys before generating recommendations across Xometry's 20 supported manufacturing processes.
Xometry also introduced a new generation of cost-prediction models beginning with computer numerical control (CNC) machining. The company said the upgraded models evaluate a broader range of manufacturing variables, including geometry, materials, surface finishes and whether a part is produced individually or as part of a larger manufacturing job.
The pricing models generate calibrated confidence levels for each CNC quote, the company said, allowing the platform to assess pricing certainty before returning an estimate. Xometry said it plans to expand the upgraded models to additional manufacturing processes supported by its Instant Quoting Engine while adding more CNC materials over time.
Supplier partners are also expected to benefit from new adaptive sourcing models, according to the company. The updated system evaluates machine dimensions, specialized manufacturing capabilities, historical quality performance and on-time delivery records when matching jobs with suppliers.
The company said live testing showed the cost-prediction models improved CNC cost-prediction accuracy by approximately 15%. Xometry also said its automated manufacturing recommendation was accepted by buyers more than 85% of the time during early deployment, with the strongest gains among new customers.
>
>
Key Takeaways
- Xometry upgrades its AI models to enhance manufacturing recommendations, pricing accuracy, and supplier matching.
- Automated recommendations accepted over 85% of the time during early deployment, showcasing user trust.
- New context-aware recommendation engine evaluates part geometry and materials for optimal manufacturing process selection.
- AI learning improves with each part quoted, creating a more interconnected intelligence layer in manufacturing.
- Xometry aims to provide customers with increasingly accurate insights for each new project submission.