DFW Sheet Metal

DFW Sheet Metal fabricates ductwork for commercial mechanical work. Sales ran in Pipedrive. Everything after a deal was approved ran on a Google Sheet and whatever people remembered about which contractors had bid a job, which estimator owned it, whether the quote had gone out, and where a job sat on the shop floor. When we went through their documentation, the problem was the gap between the sales record and the shop. Sales could not see production or delivery status without asking, the shop had no single place to see what was ready to fabricate, and delivery contacts and timelines were not visible without chasing. Leadership wanted to see bottlenecks and delivery risk before they became a missed date, and the shop team needed something they could update from an iPad in seconds.

This was a Smartsheet implementation with a Pipedrive integration. We used to implement Smartsheet as a partner, and this is that work.

The design kept Pipedrive as the system of record for sales and made Smartsheet the production execution layer. Discovery confirmed the flow from sales to production to delivery, defined the data model (jobs, batches, stages, delivery), fixed the manufacturing stages and batching rules, and set delivery contact ownership, iPad use on the floor, and who could edit what.

The build started at the front of the process. A bid request form goes to contractors. When several contractors bid the same job, the office combines them onto one row, and an auto-generated job number (with an internal sub-number for multiple jobs under one project) gives every piece of work a unique ID. Checking "ready for bid list" moves the row into a bid list where an estimator is assigned and alerted, plan links and pricing are attached, and approval records a completion date that feeds back to Pipedrive. A data sheet rolls completions up by week, month, and year, driving a dashboard and estimator reports, one per estimator showing only their pending jobs, plus a weekly view. Production sits downstream, populated by reference from the job record, with a stage per shop operation (takeoff, shop data, posted materials, plasma, coil line duct, spiral pipe, seams and brake, liner, accessories, assembly, layout, welding, and round). Views were built by workstation so a person at the plasma table sees plasma work.

Automations moved rows between sheets on status, alerted the assigned estimator when a job changed, recorded dates on completion, and were designed so that completing a stage triggers the next stage's readiness, with QA gating, delivery readiness triggers, and exception handling for blocked work. The Pipedrive integration was scoped so approved deals create jobs, deal metadata stays in sync, and delivery contacts sync one way with tagging, so operations can manage contacts without touching the CRM. An n8n workflow was designed to compute hours per stage from stage timestamps once a day and write them back to the job. Alongside the grids, we built a mockup of an application front end for the production stages, which is where the engagement went next.

Every job carries one number from bid to shop, so the office, the estimator, and the floor are talking about the same thing. Estimators are told when a job lands in their list instead of finding out later. Approval and completion dates are recorded by the system, and the weekly and monthly roll-ups exist without anyone building them by hand. The bid-to-production flow that lived in a spreadsheet and in people's heads became a set of automations with defined triggers, which made the next step possible. The client could see the process clearly enough to want it as an application rather than a set of grids. The engagement began in December 2025 and the build documentation dates from early 2026.

Given the same problem now, we would skip the grid layer and build the shop its own system from the start, holding bids, estimating, jobs, batches, shop-floor stages on a tablet, delivery, and hours per stage in one database, with Pipedrive or a built-in pipeline as a native connection and no per-seat cost for the people on the floor. That is the direction this engagement took. The Smartsheet build proved the process, and a signed change order after it began moving the system into an application of their own, with a proper backend underneath.