AI quoting for injection molders: piece price, tooling and cycle times

How AI quoting builds a molded part price#

AI quoting for injection molding builds a piece price by estimating part weight, press size and cycle time from the part model, then applying resin cost, machine rates, labor, scrap, secondary operations and packaging. Tooling is estimated alongside it from cavitation, part complexity, slides and lifters, hot runner choice and surface texture.

The software is only as good as the assumptions underneath. Machine rates, scrap allowances and cycle estimates are where molders differ from each other, and they are exactly the inputs a generic model gets wrong for your plant. Your own production history is what corrects them.

Quote inputs and the records that validate them#

Every quote input has a matching historical record that shows whether past assumptions held. Lining them up is the first job in any quoting project, before comparing vendors.

Two cautions apply. Your machine rate table must be current, because a model tuned against outdated rates will reproduce old costs with confidence. And records must be compared like for like: a cycle time logged during a mold trial or a startup is not a production cycle, so tag runs by type before comparing them to quotes.

Why cycle time is the first input to check#

Cycle time is the first input to check because it sets machine time, and machine time is a major part of most molded piece prices. Estimates start from the thickest wall and the resin's cooling behavior, but the actual cycle also depends on mold cooling design, robot takeout, insert loading and how the tool runs once it is warm.

Compare quoted cycle times with actual run cycles for every mold you can, grouped by resin family, wall thickness and part features. Consistent gaps show where the estimate logic is off, such as parts with metal inserts or glass-filled resins, and correcting them is worth more than any new software feature.

Tooling: quote it separately, amortize it deliberately#

Tooling should be quoted as its own line, then either billed on a tooling purchase order or amortized into the piece price by explicit agreement. Amortizing quietly into piece price shifts volume risk to the molder: if the program ends early, the unrecovered tool cost stays with you.

Ownership needs the same clarity. Many customers own their molds outright, with tool tags and a tooling agreement covering maintenance, storage and removal rights. Keep toolmaker quotes, build invoices, trial reports, engineering changes to the steel and maintenance logs with each tool record, because they validate future tooling estimates and answer customer questions when a mold moves.

Records to clean before you tune a quoting model#

Before you tune or train a quoting model, clean the records that connect a quote to what actually happened on the press. Where quotes were built in spreadsheets outside the ERP, gather them into one table with a consistent part key; a model trained on half the history learns half the lessons.

• 1Quote log: quote number, customer part number and revision, annual volumes, price breaks, resin, cavitation and quoted cycle.
• 2Win or loss outcome, with the reason when you know it.
• 3A link from each won quote to its tool and work orders.
• 4Actual cycle, scrap, downtime and resin usage per run from MES or press monitoring.
• 5Tool build actuals: toolmaker quote, final cost and changes during build and trials.
• 6Engineering change notices that changed the part or the tool after quoting.
• 7Customer complaints and returns tied to part numbers.

Illustrative example: a custom molder tunes its estimates#

Illustrative: a fictional custom molder runs a range of presses for appliance and industrial customers. Quotes are built in a spreadsheet template, production runs through an ERP with press monitoring, and tooling is outsourced to several toolmakers.

Before adopting an AI quoting tool, the estimating lead links several years of won quotes to actual runs. Two patterns appear: quoted cycles on parts with molded-in inserts are consistently short, and startup scrap on certain engineering resins runs above the standard allowance. Lost quotes cluster on large parts, where the molder's press rates are less competitive.

The molder corrects its cycle and scrap assumptions, sets the AI tool to flag quotes in those categories for estimator review, and starts recording a loss reason on every closed quote.

Confidentiality, ownership and licensing your quote history#

Quote history mixes records a molder owns with information it holds for others. Customer part models and drawings are usually the customer's property, customer pricing is confidential, and resin contract prices may fall under supplier confidentiality terms. Check those obligations before uploading part files to a cloud quoting tool, and read the tool's data-use terms.

The molder's own records, such as actual cycle times, scrap, process sheets and the reasoning behind quote decisions, are different. They describe how a plant runs and may be licensable once customer identities and prices are removed. SourceX begins with a metadata-only fit check, and the SourceX five-step transaction handles rights, preparation and the molder's approval before anything is delivered.