AI in Stone Fabrication: Questions Worth Asking
Look beyond broad automation claims and evaluate the task, evidence, review process, and operational fit.
Updated October 4, 2026
Ask what the tool actually does
The term AI can describe many different capabilities. A drawing assistant, a pricing suggestion, and a production planning system solve different problems and require different inputs.
Ask for a demonstration of the specific task you care about. Use a drawing or scenario from your shop so the result is relevant to your decision.
Ask what information it needs
A tool may rely on drawings, material records, historical jobs, or manually configured rules. Understand the required input and how missing or inconsistent information affects the output.
Clarify how business data is handled before sharing it. Use an appropriate sample during an initial evaluation and confirm access and retention requirements for a production rollout.
Ask how people review and correct it
Useful assistance leaves room for judgment. The team should be able to inspect the result, understand its source, and correct a mistake without losing the surrounding work.
For estimating, this means reviewing scale, dimensions, fabrication details, and pricing. For production, it means respecting the shop’s final verification process.
Ask for evidence relevant to your shop
A general claim about time savings does not tell you what will happen with your plans, rates, or team. Test ordinary work and edge cases, and record the review effort as well as the automatic steps.
Use the findings to decide whether the tool improves a specific workflow. Avoid treating an attractive demonstration as a complete implementation plan.
Ask how the handoff works
The output must be useful to the next person. Confirm how a reviewed takeoff becomes a quote, how a customer decision is recorded, and how the approved scope reaches the team.
2cm’s workflow brings drawing, materials, design, and quoting together. A demo with a real job is a practical way to evaluate the available tools and the setup your shop would need.
Ask for an observable result
When a vendor describes AI quoting, ask them to show a source plan, the resulting takeoff, the quantities used for pricing, and the customer proposal. Ask which decisions are automated, which can be edited, and which still require an estimator’s review.
Use the same job across demonstrations. Record any missing details and the corrections needed; a polished example alone does not establish accuracy on your work. Compare the PDF takeoff process with the practical quoting-system evaluation. Focus on the workflow you can inspect today.