Frequentlyasked questions

Naya AI Estimation is the best AI-powered manufacturing cost estimation tool in the world, turning a multi-week costing process into minutes. Starting from as little as a product image, sketch, or CAD file, an agentic AI architecture orchestrates specialized agents for manufacturing engineering, cost research, value engineering, simulation, material selection, tariffs and duties, and logistics to produce a per-unit and landed cost broken across materials and full BOM, labor, process costs, tooling, overhead, margin, packaging, freight, and duties — each with a confidence score and source citations. The tool supports the full product lifecycle from early-stage feasibility through final production costing.

Naya supports over 200 different file formats. The estimation tool accepts a wide range of inputs at varying levels of detail. At minimum, you can start with a product image or sketch. For more accurate results, you can upload 3D CAD files (STEP, OBJ, FBX, STL, PRT, and more), technical drawings (PDFs), BOMs, tech packs, and supplementary documents. The system also allows you to add additional context such as target volumes, materials, manufacturing region, and special requirements. The more information you provide, the higher the confidence score of the estimate.

Visit naya.studio/ai-estimation and upload your product images (up to 5). Click 'Estimate Cost' to enter the app. The AI will ask clarifying questions and may request additional source files like BOMs, PDFs, or docs. It then generates an Action Plan with step-by-step analysis. You can provide feedback at any step and rerun. The final report includes a per-unit estimate, confidence score, cost summary (labor, material, overhead, etc.), detailed step-by-step breakdown, assumptions, and data source links. You can then save the estimate in Naya’s Workflow AI product.

Yes. The estimation tool has sample SKUs on the landing page that let you see how the tool works with pre-built product examples. Samples fill your inputs so you can review the setup before generating an estimate. Enterprise prospects typically start with an out-of-the-box evaluation period during which the costing team uses the tool to build comfort, then proceed into a low-lift evaluation period. This engagement includes working closely with the Naya team to increase estimation accuracy to 99%, show ROI, and create a business case. Naya can also work with you to run samples and incorporate feedback.

The tool is used across a wide range of physical product categories including footwear, furniture and home goods, consumer electronics, industrial components, hardware, fashion and apparel, CPG, jewelry & luxury goods, automotive, sporting goods, architectural products, and more. The tool works best when there is established manufacturing data for the product category, though it can generate directional estimates for novel product types as well.

Industry Skills are specialized knowledge modules that tune the estimation engine for specific manufacturing domains. Skills include detailed process parameters, rate tables, material-specific data, and decision matrices. The system automatically matches the best skill for your product, and now displays match confidence percentages so you can see why a particular skill was recommended. Skills can be customized with your enterprise's tribal knowledge and proprietary data for even better accuracy.

Costing teams use the tool across several key workflows: (1) Pre-costing for early-stage feasibility checks before committing to a design direction, (2) Validating supplier quotes by comparing AI-generated breakdowns against factory submissions, (3) Running quick cost comparisons across different design approaches to enable parallel work streams, (4) Generating directional cost estimates from just a sketch or image when detailed specs aren't yet available, and (5) Scenario modeling to evaluate different manufacturing regions, volumes, and materials.

Yes. This is a key use case. While factories may not directly price-match an AI estimate, having a detailed cost breakdown gives teams leverage to ask suppliers for the same level of detail.

Both. It’s used for scratch builds and for making tweaks to existing lines (changing finishes, sizes, etc.).

The platform covers 100+ manufacturing processes spanning metal, plastics, and assembly. This includes 17 metal process skills (CNC machining, sheet metal, die casting, forging, extrusion, and more) and 7 plastics process skills (injection molding, blow molding, thermoforming, and others). A decision matrix selects the right process based on volume tier and material, with mandatory and optional secondary operations and material conflict detection.

Yes. The estimation includes configurable tariffs, duties, freight, and logistics costs in addition to the FOB (Free on Board) price. Cost labels dynamically adapt to show FOB or landed cost based on your estimation context. This is particularly important in the current tariff environment and has been validated with enterprise teams. The scenario planning functionality can then be leveraged for finding the optimal routes.

Yes. In the “Key Cost Drivers” or “Cost Summary” section, you can toggle shipping/logistics and tariffs on or off to see the landed cost.

Yes. Regional pricing accounts for labor rates, material costs, freight impacts, and lead times across different manufacturing geographies. Hourly rates are configurable per region and machine type. The scenario optimization feature can recommend Pareto-optimal manufacturing origins across countries, showing trade-offs between cost and lead time to help you find the best-cost option.

Cost breakdowns are provided for all cost drivers, including materials, components, parts, etc. The granularity is also customizable and we can provide costing at the granularity that is currently in practice in the organization.

Yes. The platform supports batch processing for parallel estimation at scale. Multiple products, parts, or assemblies can be uploaded and processed simultaneously through the 'batch run' feature. Performance characteristics scale with input volume and integration depth; specific throughput numbers for a particular deployment are mapped during the implementation period.

Yes. The platform models costs across offshore and nearshore manufacturing in all major regions, with manufacturing country, volume tier, and material/process combination modeled directly in each estimate. Supplier-type differences (contract manufacturers, tier 1 suppliers, regional configurations) are captured by configuring region-specific labor and machine rates, machine selection, overhead allocations, and freight and tariff assumptions. The Manufacturing Partners feature suggests factories and suppliers matched to the product being estimated. For enterprise deployments, your approved supplier lists, supplier rates, and proprietary supplier data can be integrated so suggestions reflect your own criteria.

Value Engineering is an analysis feature that identifies cost reduction opportunities across your product. It provides recommendations organized by category with an interactive bar chart visualizing potential savings. You can set a target price and the system will generate compliant design/estimation options to meet time or budget constraints. Value Engineering can also be rerun directly from Key Cost Drivers with automatic progress tracking, making it easy to iterate toward an optimal cost structure. Value engineering functionality improves with customer data and expert feedback.

Scenario Planning allows you to compare different manufacturing scenarios side by side. You can model variations in manufacturing country, volume, material, and other parameters. The AI-powered scenario optimization recommends Pareto-optimal manufacturing origins with landed cost comparisons. The view uses a sortable table with country selection, making it easy to compare scenarios A, B, C, and D. Users have requested and we support tooling amortization, CapEx, first-year forecasts, and MSRP analysis within scenario planning.

Competitor Checkup (also called Competitive Analysis) lets you benchmark your product's manufacturing costs against similar products in the market. Users can input competing products to get an early check on what prices need to be achieved. Enterprise customers have found the competitor analysis section particularly valuable for understanding what similar products cost in the marketplace.

Product mode estimates the full cost of an assembled product, including all components, assembly, and packaging. Parts mode estimates individual component costs in detail, including specific manufacturing process parameters like machine tonnage, cycle times, and material densities. Both modes include the same analysis features (value engineering, scenario planning, etc.). For sub-assemblies, you can start with parts mode for individual components and scale up to full product estimation.

Estimates cover the cost to design and manufacture a product, including materials and full BOM, labor, process costs, tooling, overhead, margin, packaging, freight, and duties, rather than ongoing costs incurred after a unit is in the field.

What the platform does model today is how durability and longevity requirements change initial cost. Specifying attributes upfront, such as weatherproofing, impact or tamper resistance, or a target service life, lets those requirements be priced into the estimate. The cost of different durability specifications can then be compared side by side.

Yes. The tool works with just a 2D fabrication drawing, without any 3D renderings or visual images. The system will extract what information it can from the drawing to generate an estimate. Providing more detail (dimensions, materials, tolerances) in the drawing will improve accuracy.

The platform supports 70+ CAD and 3D model file formats for automated cost modeling, with the strongest geometry extraction on neutral mechanical CAD.

Native CAD parsing (BREP):

STEP / STP (.step, .stp)

IGES (.iges, .igs)

Parasolid (.prt)

Also supported for visual / AI-assisted analysis:

GLB / GLTF, OBJ, STL

Fusion 360 (.f3d) — preview + add-in workflow

Onshape — live cloud import via OAuth

For mesh-only formats (STL / OBJ / GLB) we rely on visual analysis and mesh metadata; parametric feature extraction is strongest with STEP / IGES.

If a file type from your workflow is not supported, let us know at info@naya.studio.

From neutral BREP CAD formats (STEP, IGES, Parasolid), the platform extracts dimensions, bounding box, volume, surface area, face/edge/body counts and surface-type breakdown (planar, cylindrical, conical, B-spline), holes (diameter, depth, through/blind), fillets (radius, count), and wall thickness (min/max/avg) — feeding process selection and machining complexity. Sheet-metal parts are auto-detected and unfolded with flat-pattern dimensions, cut perimeter, bend count, thickness, nesting/utilization, pierce counts, and a downloadable DXF. For mesh-only formats (STL, OBJ, GLB), the platform relies on visual analysis and mesh metadata. Parametric feature extraction is strongest with STEP and IGES.

Out of the box, the tool achieves approximately 85% accuracy using public data sources for material pricing and manufacturing rates. Accuracy improves significantly when enterprises integrate their own supplier rates, factory pricing, and proprietary material databases. During the 2-4 week evaluation period, accuracy increases to 99% for one product type or line demonstrating the business case. Companies may then enter paid contract agreements to cover any remaining product lines.

After our implementation period this is 99% or higher. Many teams report that our accuracy is even higher than that of their estimators over a testing period.

The confidence score reflects how complete and reliable the estimation inputs are along with the model’s familiarity with your specific quoting process. A lower score typically means the tool is working with limited information (e.g., just an image); a higher score indicates detailed inputs like CAD files, material specs, and volume targets were provided. The system identifies missing information and prompts you with specific additions that would meaningfully improve confidence (for example: 'If you add details X and Y, your confidence score will increase by 10%'). Working with Naya on an enterprise implementation is the best path to consistently high confidence scores.

For enterprise users, the best way to increase accuracy is to work with us to integrate your proprietary supplier rates and material databases. Schedule a call to learn more at: https://naya.studio/request-demo

Additional strategies improve accuracy: (1) Upload multiple input types — combining images, CAD files, and documents yields better results than any single input. (2) Provide additional context about target volumes, materials, and manufacturing region. (3) Select the correct industry skill for your product type. (4) Review and edit key parameters after the initial run.

Yes. Every field in the costing summary is fully editable, including overhead, margin, tooling, labor rates, and material costs. An editable Key Numbers panel lets you adjust cost-related parameters directly. The system learns from your edits over time to improve accuracy for future estimations. You can also rerun individual steps or the entire estimation with modified inputs.

Yes. Estimation reports can be shared via a shared estimation viewer link, which includes manufacturing details, BOM, and technical tabs. Reports can also be exported as PDFs and saved as Naya projects within Naya’s Workflow AI tool. Shared reports now display which industry skills were used during the estimation. All output fields in the costing summary (overhead, margin, tooling, etc.) are fully editable, so your team can refine estimates collaboratively.

Estimations can be saved directly into Naya's Workflow AI tool as Naya projects (which includes all images, details, documents, action plans, report PDFs, and source links in an organized project). Reports can be exported as PDFs. Saved reports lock pricing at the time of save, while live projects auto-update with current market data.

Enterprise customers can integrate their own supplier rates, material databases, factory pricing, and proprietary process data to significantly improve estimation accuracy beyond the public data baseline. The system supports a two-phase approach: first, teams evaluate using the out-of-the-box version to build comfort, then move into refinement with actual data sets and integrations—typically through a 2-4 week low-lift period. The system then learns from expert edits over time, generating increasingly accurate results. Company data is not used to train models. Skills learn and improve over time within your organization and are not shared across organizations.

Naya estimation AI can use existing data systems to increase accuracy and consistency of results. Any historical data in these systems can also be used to train the skills for your organization.

We continuously update our different pricing sources through APIs, third party data providers and also scrapers. We can also plug into your data sets which have live feeds. We make sure that all cost drivers have the most updated data sources when running an estimation.

Naya is part of the PTC ecosystem. We natively support the file formats commonly used in Windchill workflows, including STEP, IGES, Parasolid, JT, and Creo, and we have a live cloud integration with Onshape via OAuth one-click import.

If used by your organization, the Windchill connector itself is built and tuned during the implementation period to fit your company's specific environment. Our APIs are robust and compatible with PTC systems, and a connector to any PLM can be stood up in a matter of days. Connection approach (CSV, API, or custom connector) and data flows (BOMs, part data, materials, design files, historical cost data) are scoped at that stage.

Naya is SOC 2 Type I and Type II approved through Vanta. The report is available on request, and additional security documentation can be shared as needed. More information at https://app.vanta.com/naya.studio/trust/lz494cq8y8mqf9bzt21gth

We take data security and IP very seriously as this is a major consideration for our enterprise customers.

We are a SOC2 ready enterprise with best practices in cybersecurity. We also deploy our production solutions in firewalled and enterprise isolated networks so data never leaves servers that are not controlled by enterprises. We can also deploy on-prem on request.

Our databases are isolated by tenant, encrypted at rest and in transit. We have strict access control setup that can be configured with the enterprise IT teams. We have gone through the IT and cybersecurity process with all our enterprise customers' IT teams and we can start early and expedite the process.

Customer data is strictly isolated per tenant, encrypted at rest and in transit, with configurable access controls set up alongside your IT team.

Product designs, BOMs, and pricing data are never used to train foundation models and are never shared across accounts. Skills (the customer-specific knowledge layer) tune and improve over time within your organization only and are not shared with other customers.

Naya does not train or fine-tune AI models. The platform uses third-party foundation models created and trained by Anthropic (Claude) and Google (Gemini), accessed via API under commercial terms that prohibit use of customer data for model training. Training data sourcing for these models is governed by each provider's published documentation.

Naya's domain expertise is applied through proprietary prompting, structured knowledge (manufacturing rates, material libraries, component glossaries), and expert-in-the-loop validation, none of which involves training on customer data.

No. Customer inputs are processed at inference time only, and neither usage data nor outputs are used to train or fine-tune any model; the models are third-party and accessed via API only. Outputs are stored in a dedicated per-customer environment and used solely for that customer. Skills, the customer-specific knowledge layer, can improve from your team's edits within your organization's environment only, which does not involve model training.

No. Customer inputs and outputs are used exclusively to deliver the contracted service. Any other use requires explicit written permission from the data owner.

The customer retains all rights to its inputs and outputs. Subprocessors receive processing access only, with no rights granted or transferred.

No. Outputs may be used freely for your business purposes.

Yes. The service is governed by a written agreement. Data use and AI commitments are contractual and cannot be changed unilaterally.

The default deployment is a separate cloud instance per company. On-premise deployment is also available for organizations with strict security requirements.

Naya supports NDA processes and IT security evaluations as standard parts of the enterprise onboarding workflow. Naya can start the IT and cybersecurity process early in the engagement to expedite contracting. Customer references are available on request.

A current list of sub-processors is available on request, and we notify customers in advance of any material changes.

Naya supports SSO via SAML and OIDC through WorkOS, our enterprise authentication provider, so you can connect your existing identity provider (Okta, Azure AD, Google Workspace, and others). This is paired with role-based access control to scope what each user can view and edit, plus audit logging of key actions. These can be configured alongside your IT team during onboarding.

Intended use:

• Design to Cost Time Savings. Accelerate costing and design to cost process, saving hours of time.
• Procurement Cost Savings. Provide procurement teams with data and insights to have more informed conversations with suppliers and select suppliers sooner.
• Increase Revenues. Enabling teams to get more products to market faster by making better design and cost decisions upfront with target price in mind.
• Higher Quality Products. Better value engineering, material choices, sustainable selections and design decisions that lead to overall products that are more loved by customers.

How it should not be used: Estimates are not binding quotes, so an estimate should not be the sole basis for an irreversible financial or contractual commitment without human validation. Estimates built on limited input, such as an image alone from an early stage design, may carry a lower confidence score (~90%) than estimates from later stage designs (~99%) and should be treated as directional until richer inputs are added.

Human monitoring:

Human monitoring is concentrated in the deployment phase and decreases significantly as confidence and accuracy are systematically established throughout deployment. In a typical engagement, cost models are built iteratively over the first weeks, moving from single-process parts to full products, with weekly subject matter expert review calls in which estimates are walked through and refined. Accuracy is validated stage by stage, and each product line is signed off by its SMEs before coverage is locked. Integration and data-flow architecture are then designed and approved with the customer's IT and security owners, followed by a pilot phase in which the organization's own users run real costing exercises with training and structured feedback. A final review measures accuracy lift and time savings.

After deployment, routine oversight requirements drop considerably. Controls stay in place if a reviewer wants them; rate tables and cost parameters remain editable, expert edits continue to refine the organization's Skills, and confidence scores and source links appear on every estimate.

Naya runs on Google Cloud Platform (GCP), using Cloud Run, Firestore/Datastore, Cloud Storage, and Secret Manager, with MongoDB Atlas for database hosting, all in GCP's SOC 2-compliant data centers. Infrastructure is currently hosted in the US, and both GCP and MongoDB Atlas support regional data residency, so data can be provisioned in specific geographies including the EU. The default enterprise deployment is a separate cloud instance per company, with on-premise available for stricter requirements. Data is encrypted at rest (AES-256) and in transit (TLS 1.2+), isolated by tenant, and protected by least-privilege access controls and continuous monitoring. Architecture documentation and our SOC 2 report are available under NDA on request. See Naya's Trust Center: https://app.vanta.com/naya.studio/trust/lz494cq8y8mqf9bzt21gth

Naya Studio processes and stores limited personally identifiable information (PII). User authentication and identity data are managed through WorkOS, our authentication provider, with some user data (such as account details) also stored in our own database.

All PII is classified as Confidential under our Data Management Policy, encrypted in transit (TLS 1.2+) and at rest (AES-256), access-controlled on a least-privilege basis, and retained only as long as needed for a legitimate business purpose.

Our Information Security Roles and Responsibilities policy assigns ownership for security across the organization, from executive accountability down to individual engineering and operations roles. See Naya's Trust Center for more information.

Our Incident Response Plan defines detection, triage, containment, and communication procedures. Customers are notified within 24 hours of a confirmed security incident, with root cause analysis and a remediation plan delivered within 5 business days. See Naya's Trust Center for more information: https://app.vanta.com/naya.studio/trust/lz494cq8y8mqf9bzt21gth

Our Data Management Policy covers classification, handling, retention, and secure disposal. Data is encrypted at rest (AES-256) and in transit (TLS 1.2+), isolated by tenant, and never used to train models or shared across accounts. On termination, data is deleted or returned within 30 days with certification. See Naya's Trust Center for more information: https://app.vanta.com/naya.studio/trust/lz494cq8y8mqf9bzt21gth

Naya maintains documented Business Continuity and Disaster Recovery plans covering our Google Cloud infrastructure. See Naya's Trust Center for more information: https://app.vanta.com/naya.studio/trust/lz494cq8y8mqf9bzt21gth

Our Asset Management Policy tracks ownership, inventory, and lifecycle for all hardware, software, and cloud resources that touch customer data, including secure decommissioning. See Naya's Trust Center for more information: https://app.vanta.com/naya.studio/trust/lz494cq8y8mqf9bzt21gth

Our Third-Party Management Policy requires security review of vendors before onboarding and ongoing monitoring thereafter, scaled to the sensitivity of the data involved. See Naya's Trust Center for more information: https://app.vanta.com/naya.studio/trust/lz494cq8y8mqf9bzt21gth

Naya maintains a cyber liability insurance policy. A certificate of insurance and policy details are available on request. See Naya's Trust Center for more information: https://app.vanta.com/naya.studio/trust/lz494cq8y8mqf9bzt21gth

Pricing ranges depending on organization size, usage volume, integrations, and customization requirements. The product has demonstrated massive ROI, with payback periods as short as a few weeks of implementation. Enterprise engagements typically begin with a POC or evaluation period. Bundling discounts are available for multiple Naya products.

The best way to determine ROI for your organization is to schedule a call with us: https://naya.studio/request-demo

We also have an ROI calculator here: https://naya.studio/estimation#roi-calculator

Some questions to consider:

How many times per month/year are costing requests received by the team?

How many team members are supporting the costing function, and at what capacity?

How long does it take to generate a cost on average? Helpful if this is broken out by type of product or stage of design

How accurate are the estimates? Are there steps of estimation that are more lengthy or likely to be less accurate?

Is there a wait time to get an estimate? If so, how large is the queue, or how long does it take to get back to other teams?

Does the costing team require feedback from factories or manufacturing or other partners?

How many back-and-forths are there between the design and estimation teams?

How often are designs value-engineered or edited for cost?

How many products don’t go to market because they were too expensive or not designed within the target budget?

The Sustainability AI product supports circular economy modeling (recycled and alternative materials, remanufactured components, design-for-disassembly), and regional sustainability cost factors like tariffs, duties, and freight by manufacturing country. Carbon taxes, Extended Producer Responsibility, and packaging regulation modeling can be customized for enterprise deployments.

Yes. Naya offers a dedicated Sustainability AI product that assesses carbon, water, plastic, and energy footprints across materials, manufacturing, and transport, from an image all the way to a final BOM. Sustainability AI is designed to run alongside Estimation AI on the same inputs, so cost and environmental outcomes are weighed together rather than handled as separate workstreams. There is also a Green Packaging estimation mode.

Naya models regional cost factors including tariffs, duties, freight, and logistics by manufacturing country. Sustainability AI assesses carbon, water, plastic, and energy footprints. Modeling of specific regulatory regimes, including carbon taxes, Extended Producer Responsibility, packaging regulations, can be customized for enterprise organizations as part of implementation.

Naya has 7+ years of cost modeling and should-cost experience, with the Estimation AI product itself in active development for over 7 years. The work originated as research at Harvard's Design and Engineering School, where the first generation of the AI was trained specifically for furniture. The platform has since expanded across 11+ verticals and is now used by over 1,000 companies.

The headquarters is in New York, but the team is spread across the US, India, and Europe.

Naya software has been recognized with the iF Design Award 2024 as Winner in the User Experience (Product UX) category. Beyond recognition for Naya itself, products designed and developed on Naya have been recognized across leading industry awards including iF Design Award, European Product Design Award, IDA Design Awards, Red Dot Award, SF Design Week, Milan Design Week, and IDEA Awards. Naya is SOC 2 Type I and Type II certified through Vanta. Naya has also been featured in Autodesk, Onshape by PTC, FastCo, Design Burger, USC, and Metropolis press.

Naya's enterprise stack is built around strategic alliances with major CAD, design, and product development platforms — including PTC, Autodesk, and Trimble, each with multiple integrations. ERP and other enterprise systems are supported through flexible connection paths (CSV, API, or custom ERP connectors), scoped per customer. The platform supports 200+ integrations and file types in total.

Yes. Customer feedback has been core to Naya's product development from day one, with the platform built on 10,000+ user insights. The advisory board includes senior leaders from leading product brands, top design schools, and Fortune 500 companies — including the CFO of Figma and the Co-Founder of Pinterest, alongside academic advisors from MIT Sloan and USC. A Design Advisory Board of core customers provides frequent product feedback.