AI-assisted quoting in skilled trades: from enquiry to draft

Bring emails, photographs, site measurements, drawings and conversation notes together under one record identifier. Record the applicable drawing version and the source of each detail. The application can organise this material into a summary of the work, location, materials, dimensions, specification and requested date. Missing information becomes questions for the person handling the enquiry.

For example, a fitted wardrobe enquiry includes a photograph and approximate dimensions. The customer requests a particular finish, but details of the installation setting are missing. The first useful output is an organised enquiry showing these open points. A finished quotation would merely carry the uncertainty forward. Decide which information is sufficient for an initial assessment and which must be available before final costing.

Mittelstand-Digital Zentrum Hannover describes an INOMETA feasibility study: from 500 assemblies per class, time predictions in the historical dataset outperformed the previous method. Drawing analysis was prototypical; transfer into industrial use remained further work. This is a study result, not a universal dataset minimum for trade businesses. Source checked on 10 September 2026.

In your own workflow, make drawing features, time estimates and pricing separately reviewable. Predicting production time requires comparable jobs with actual recorded times. Current material prices, labour rates and overheads belong in the subsequent costing; changed production methods need reassessment.

• Mittelstand-Digital Zentrum Hannover: INOMETA feasibility study, checked 10 September 2026 (German)

In the DigiCast episode of 6 January 2025, Matthias Brack described an AI application for assessing customer enquiries at Brack Wintergarten. He reported that the business had collected structured enquiry data for more than six years. A commissioned model used that history to estimate the likelihood of a subsequent order and support the consultation process.

This account concerns sales management. It does not establish automated technical costing from drawings. Before considering a similar model, examine what its output would change: callback order, additional questions or the next consultation step. Missing information should not become a confident judgement about the customer. Define how new information and differing assessments enter the process.

• Mittelstand-Digital Zentrum Handwerk: DigiCast episode 40 with Matthias Brack, 6 January 2025 (German)

Technical review checks the specification, quantities and necessary work. Commercial review examines prices, overheads and scope. In a small business, one person may carry out both. The essential point is that the review uses the actual data and is recorded against the quotation.

A revised customer drawing may affect several line items. Show what has changed since the previous version and which calculations need another review. Keep the sent version with its supporting information. Later changes to master data should remain traceable without silently rewriting an earlier quotation.

• Does every line item match the enquiry and current drawing?
• Are quantities, units, variants and rounding correct?
• Are travel, measurements, disposal and customer preparations covered?
• Have open assumptions been resolved or made explicit in the quotation?
• Are approved prices, terms and dates carried over unchanged?
• Can the business establish which version was approved and sent, and by whom?