The Challenge
A California-based custom manufacturer with 40,000 SKUs and a 70-year operating history came to Remix Partners after hitting the ceiling of what their internal team could build. Their RFQ process was deeply manual — engineers printed specs from legacy software, manually re-entered data into a second system, advanced a bill of materials by hand, and generated quotes through constant back-and-forth with customers. Their target of 48-hour RFQ turnarounds was almost never achieved.
What They Built
Remix Partners ran their GenAI Kickstart process — interviews with engineers, operations leads, and the executive team to map workflows and identify the highest-leverage friction points. Their first move was converting the output of a legacy software system into JSON, which unlocked downstream automation. Over 3.5 weeks they built a Claude Code-based workbench that: ingests an RFQ, compares it against all prior parts the company has made, generates a draft bill of materials, notifies team members, converts the draft BOM into a price quote, and takes a first pass at production drawings. Custom AI validation flags specification conflicts before they reach the floor. Engineers review and refine a 90% complete draft rather than starting from scratch.
Remix Partners started with their GenAI Kickstart process — structured discovery interviews with engineers, operations leads, and the executive team to map the quoting workflow and identify the highest-leverage friction points. One finding stood out: the legacy software system produced proprietary output that was unreadable by other systems, making automation impossible until that was solved.
Their first move was converting the legacy system's output into JSON — a foundational unlock that made every downstream automation step possible. From there, over 3.5 weeks, they built a Claude Code workbench covering the entire RFQ-to-production workflow: the system ingests an incoming RFQ, compares it against all 40,000 prior parts the company has manufactured, generates a draft bill of materials, notifies team members, converts the draft BOM into a price quote, and takes a first pass at production drawings. Custom AI validation flags specification conflicts before they reach the manufacturing floor.
Engineers no longer start from scratch — they review and refine a 90% complete draft, freeing experienced staff from routine documentation for the complex, high-value designs where real competitive differentiation lives. Projected outcome: 60–70% reduction in engineering time per customer order.
AI Role
Claude Code powers the AI workbench that processes inbound RFQs end-to-end. It ingests a request, compares it against the full 40,000-item catalog of prior parts, generates a draft bill of materials, converts it to a customer-ready price quote, takes a first pass at production drawings, and flags specification conflicts — all before a human engineer reviews the output.
Infrastructure
Legacy proprietary software system (existing, output converted to JSON) • Parts catalog database (40,000 SKUs, ingested for comparison)
Integration Points
Legacy system output → JSON conversion layer (foundational data unlock) • JSON data → Claude Code workbench (RFQ ingestion and processing) • Parts catalog → BOM generation module (historical parts comparison) • BOM → pricing logic → production drawing generator (sequential pipeline)
CEOs, COOs, and CFOs at companies up to 500 employees (with expansion into mid-market and Fortune 500) across professional services, manufacturing, life sciences, healthcare, VC/PE, and nonprofits — specifically leaders who know AI matters but are overwhelmed by the noise and need a trusted partner to cut through it and deliver real results, not demos.
