Quote time cut down from weeks to days for CoPack, a specialty print manufacturer

The problem

CoPack, a specialty print manufacturer, had RFQs come over WhatsApp, email, and calls. Account executives passed on client requests to the quotation team unstructured, varied in detail, and easy to lose. Quotation also was one-off calculation with no clear process.

The bottleneck was the time to turn an inbound client message into a quote a customer could actually act on: 1-2 weeks of back-and-forth between clients, account execs, and the quote team on every RFQ.

Step 1: Workflow Diagnostic

We mapped the full RFQ-to-quotation flow end-to-end: inbound channels, information gaps, the implicit pricing rules, edge cases on materials and finishes.

The more important output of the diagnostic was the separation of what was deterministic (structured information you can capture in a form) from what was judgement (pricing decisions, capacity calls, exception handling). That distinction is what made the build actually useful, instead of just another tool gathering dust. As the founder put it: breaking the problem into component parts made it feel manageable instead of another massive IT project.

Step 2: Custom RFQ intake form

A purpose-built RFQ intake form that captures the right information the first time (material, dimensions, run size, finishes, delivery target) in the shape the quotation model needs. The form sits at the front of their pipeline, so by the time an RFQ reaches the quote owner, it's already structured.

The effect cuts both ways: customers get faster quotes, and the founder gets to spend his time on the actual judgement work (pricing, capacity, material choices) instead of chasing basic details over WhatsApp.

Step 3: 1:1 AI Coaching

Beyond the workflow, I worked with the CoPack leadership team on their AI foundations: current trends and the concepts that actually matter for running a business like theirs.

Topics covered included personal AI agents, deterministic vs judgement-based workflows, knowledge graphs for AI memory, and how to think about where automation fits versus where it doesn't.

The real outcome: after the sessions, the founder was confident enough to start building his own AI bots, and he hasn't stopped since. He's no longer waiting on outside help to ship new automations.

The outcome

• • Quote turnaround dropped from 1-2 weeks to 2-3 days
• • Founder saves ~2 hours/week on reviewing unfinished RFQs
• • Quotation model externalized, no longer locked in one head
• • Founder's time freed up for the judgement work that actually moves quotes
• • Founder now builds and ships his own AI bots, no longer waiting on outside help

I always thought our factory needed a massive IT transformation to do anything useful. But we focused heavily on sales, marketing & pricing, that's what let us scale digitally first. They acted as a siloed area to work first that had limited downsides.Director, CoPack

What was used

Workflow Diagnostic · Custom RFQ intake form (Next.js + Supabase) · Claude · Obsidian for AI memory · 1:1 AI Coaching sessions.