Configure-to-Order Quoting for Valve and Pump Manufacturers

Why CTO Quotes Stall in Valve and Pump Shops

A ball valve or end-suction pump might have thousands of valid combinations: body material (carbon steel, 316 stainless, duplex, bronze), end connection (flanged, threaded, socket weld, butt weld), pressure class, seat and seal material, stem packing, actuator type, and service conditions such as sour service or oxygen cleaning. The sales engineer gets a request, checks a catalog, emails engineering about an unusual combination, waits for a lead-time answer from purchasing on castings or bar stock, and then keys a price from a spreadsheet. Each hand-off adds a day.

Quotes with any non-standard element commonly take three to ten working days, and the customer has often already ordered from whoever answered first.

The Rules a Configuration Must Respect

A configurator is only as good as the rule set behind it. In valves and pumps those rules are specific and testable, which is why they can be captured rather than left as tribal knowledge.

• Pressure-temperature limits by body material and class, for example the ratings tables in ASME B16.34 for valves
• Material and service restrictions such as NACE MR0175 / ISO 15156 limits for sour service or oxygen-clean prep
• Pump hydraulics: a selected impeller trim must meet the duty point on the published curve, with NPSH margin
• Compatibility of seat, seal and packing materials with the media and temperature
• Mandatory add-ons: PMI, hydrostatic or seat test per API 598, third-party witness, certification level

From Configuration to Price and Lead Time

Once a configuration is valid, the price is a roll-up: base assembly, option deltas, material surcharges, castings or bar stock at current cost, machining and assembly hours, test labor, packaging, and any outside processing such as coating or heat treat. The lead time is the longer of long-lead purchased items (castings, actuators, mechanical seals) and the shop's open capacity on machining and the test stand. Shops that quote from a flat price book miss both: the price book lags alloy and casting increases, and the lead time ignores that the test stand is booked for three weeks.

The quote should read live standard costs, open purchase order lead times and work center load from the ERP, not a copy of them.

Where AI Helps, and What It Should Not Do

AI is useful at the front and back of the process rather than as the pricing engine. At the front, a model reads an inquiry email, data sheet or spec sheet and proposes the configuration fields: size, class, materials, trim, test level, and flags anything it could not determine. At the back, it checks the draft against past orders for the same family, surfaces actual build hours versus quoted, and drafts the cover letter with exceptions to the customer's spec.

The valid-configuration rules and price roll-up should stay deterministic, in a rules engine or ERP configurator, so a wrong answer is traceable. A human sales engineer approves anything that leaves the building.

When This Is Not Worth Building

If your catalog is a few dozen stock models, you ship from inventory, and quotes are already same-day from a price list, a configurator and AI layer will not pay back. The same is true if most of your volume is true engineered-to-order work such as custom API 610 pumps, where each job needs an engineer's judgment and the delay is in engineering, not in assembling a configuration. The sweet spot is a family with real option depth, a repeat customer base, a quote queue longer than two days, and a rules set an experienced engineer can state in a few pages.