Threekit

Threekit is an AI furniture finder tool. An AI-native CPQ for manufacturers that turns plain-language needs, photos or spec sheets into valid configurations with rule checks, price book pricing and quotes, usable by sales teams, dealer networks and website visitors.

Threekit is a configure, price, quote platform aimed squarely at manufacturers with complex catalogs. The pitch is a shift in starting point: instead of a rep hunting a part number and clicking through option trees, the buyer describes what they need in plain language, a photo, a voice note or a spec sheet, and the system begins building a valid configuration from there. The agent asks follow-up questions to fill gaps, then the rules and pricing engines check every selection and produce the number. In daily use it covers the CPQ fundamentals manufacturers expect: product catalogs and attributes, compatibility constraints applied during configuration, pricing driven by your own price books and logic, quote PDFs with line items, locking and revision, role-based price visibility for dealers versus internal teams, and handoff of the final configuration as JSON or XML. When requirements change mid-build, dependent options update and the revised build is re-checked, so a different motor or material does not send the quote back to the beginning. Good, better, best comparisons let reps discuss trade-offs with the buyer. Deployment can go two ways. Teams can replace an incumbent quoting system outright, or run Threekit in front of an existing CPQ and let an integrator wire the handoff to downstream systems. The vendor positions itself as faster to stand up than legacy CPQ, and lists Salesforce, Oracle ERP Cloud, SAP, Infor, Epicor and HubSpot as platforms it sits alongside. The homepage references case studies with Sloan and Andersen rather than publishing rates. The fit is mid-market and enterprise manufacturers with configurable product lines, dealer networks and multi-role pricing, plus websites that need to capture qualified requirements before a human gets involved. It is not a self-serve tool: pilots are run with vendor support, and scope depends on how complete your catalog data and rules already are.