← Back to Field Notes

March 28, 2026 · The Bloomfield Team

Quoting for Make-to-Order vs. Make-to-Stock: Where AI Fits in Each

A make-to-order (MTO) job shop quotes every part from scratch. Each RFQ represents a new geometry, a new material combination, a new set of tolerances, and a new cost estimate built from the ground up. The estimator's day is spent building prices for work the shop may never have done before.

A make-to-stock (MTS) manufacturer quotes differently. The products are defined. The costs are known from years of production history. The quoting challenge is not "how much does this part cost to make?" but rather "what price captures the right margin given current material costs, production capacity utilization, volume commitments, and competitive pricing in the market?" The MTS estimator is a pricing strategist, not a cost estimator.

AI serves both environments. The tools, the data they use, and the decisions they support have almost nothing in common.

The MTO Quoting Challenge

A make-to-order manufacturer processing 40 to 60 RFQs per month faces a new problem with each request. The RFQ arrives with a drawing, a material specification, a quantity, and a delivery requirement. The estimator must determine the manufacturing process, the sequence of operations, the setup and cycle times for each operation, the material cost at the specified quantity, any secondary operations like heat treatment or surface finishing, inspection requirements, and the risk factors that could cause the actual cost to exceed the estimate.

The critical input for all of these determinations is the shop's own production history. A part with 6061 aluminum, 12 features requiring 3-axis milling, and tolerances in the +/-0.005" range has been quoted and produced by the shop dozens of times, even if this specific part has never been seen before. The setup times, cycle times, and cost outcomes from those similar past jobs contain the patterns that predict what this new job will cost.

Without an AI tool, the estimator accesses this information through memory. They recall the job from two years ago that had similar features and check the ERP to see what it actually cost. They remember that 6061 in that size range was running about $3.20 per pound last time they checked. They estimate setup at 3 hours based on their experience with similar fixture requirements.

With an AI tool, the estimator opens the RFQ and the system immediately presents the five most comparable past jobs from the shop's complete history, ranked by geometric similarity, material match, and tolerance requirements. Each past job shows the quoted price, the actual cost, the variance, setup times, cycle times, and any production notes. Material pricing is current, pulled from the most recent supplier quotes on file. The research phase that took hours now takes minutes.

AI for MTO Quoting

The AI tool built for an MTO environment needs three core capabilities.

Job similarity matching. Given the parameters of a new RFQ (material, dimensions, feature count and type, tolerance range, quantity), the system searches the historical job database and ranks past jobs by similarity. The matching algorithm must understand manufacturing context: a 12" x 8" x 3" aluminum block with 6 threaded holes and a flatness call of 0.001" is more similar to a 10" x 7" x 2.5" aluminum block with 5 threaded holes and a flatness call of 0.002" than it is to a 12" x 8" x 3" steel block with no threaded holes and no flatness requirement, even though the raw dimensions of the second comparison are an exact match. Manufacturing similarity is feature-driven and process-driven, not dimension-driven.

Cost estimation support. The system presents historical cost data for the matched jobs in a format that supports the estimator's pricing decision. Quoted cost, actual cost, and the variance between them for each matched job. Material cost as a percentage of total cost, broken out by material type and size. Setup time actuals versus estimates, highlighting jobs where setup took substantially longer than expected and the reasons why (when documented). Cycle time per feature type, allowing the estimator to build a bottom-up estimate from feature-level data when no sufficiently similar whole-part match exists.

Risk flagging. The system identifies risk factors based on historical patterns. Tolerances that have historically caused quality issues on similar materials. Customer-specific requirements that differ from standard practice. Geometries that have resulted in setup time overruns. Material specifications that have historically had long lead times or price volatility. Each flag includes the historical data behind it, so the estimator can decide whether to adjust the quote or accept the risk.

The data requirements for this tool are straightforward: the ERP job history (3 to 5 years of completed jobs with cost data), the quoting files or system where past quotes are stored, and ideally the quality records that document which jobs had issues and why. The ERP holds most of this data but was never designed to surface it in the format the estimator needs.

The MTS Pricing Challenge

A make-to-stock manufacturer does not quote individual parts the way an MTO shop does. The products exist in a catalog. The manufacturing costs are established through years of production data. The cost to produce part number 4420-B is known to the penny, updated quarterly based on actual production results.

The pricing challenge in MTS is strategic. What price maximizes margin while maintaining market position? That question depends on variables that change constantly: raw material costs, production capacity utilization, inventory carrying costs, competitor pricing, customer volume commitments, and demand forecasting accuracy.

Consider a manufacturer of standard hydraulic fittings producing 200 SKUs, each with an established production cost. The pricing manager reviews prices quarterly. For each SKU, they consider the current material cost (brass, steel, or stainless, which have different volatility profiles), the production cost based on the most recent production run, the current inventory level (overstock means they can afford to be aggressive on price; understock means they need to protect margin while managing allocation), competitor pricing from the most recent market survey, and the volume-discount tiers offered to distribution customers.

Manual quarterly price reviews at this scale take two to three weeks. The pricing manager works through 200 SKUs, pulls material costs from the purchasing system, pulls production costs from the ERP, checks inventory levels, references competitor pricing data that may be weeks old, and makes pricing decisions one SKU at a time. By the time the review is complete, the material costs at the beginning of the review may have changed.

AI for MTS Pricing

The AI tool built for an MTS environment needs a fundamentally different set of capabilities than the MTO quoting tool.

Dynamic cost modeling. The system maintains a real-time cost model for every SKU based on current material costs, actual production data from the most recent runs, and overhead allocation at current capacity utilization. When material costs change, the cost model updates automatically, and the SKUs affected are flagged for pricing review. The pricing manager does not have to hunt for cost changes. The system surfaces them.

Price optimization. For each SKU, the system analyzes the relationship between price point, order volume, and margin over the historical sales data. It identifies the price elasticity for each product: how much does order volume change for a given price change? For commodity fittings where the market has five competitors, the elasticity is high and pricing latitude is narrow. For specialty fittings where the manufacturer has a proprietary design or a dominant market position, the elasticity is lower and pricing latitude is wider. The AI tool quantifies this for each SKU based on actual sales data, replacing the pricing manager's intuition with data-supported analysis.

Competitive intelligence integration. The system ingests competitor pricing data from whatever sources are available: distribution catalog feeds, customer-reported pricing, market survey data. For each SKU where competitive data exists, the system shows the manufacturer's price relative to the market and flags situations where the manufacturer's price is either well above market (risk of volume loss) or well below market (margin opportunity). This comparison is most valuable when it is tracked over time, showing how the competitive position has shifted quarter to quarter.

Volume discount optimization. Distribution customers negotiate volume-based pricing tiers. The AI tool analyzes each customer's actual purchase history against their contracted tier, identifies customers who consistently buy above or below their tier, and recommends tier adjustments at the next contract review. A distribution customer who has been buying at the 5,000-unit tier but is actually averaging 8,200 units per month should be moved to the 10,000-unit tier with a lower per-unit price, because the higher volume commitment provides better production planning visibility and the lower price still delivers acceptable margin at the higher volume.

The Hybrid Manufacturer

Many manufacturers operate in both modes simultaneously. A manufacturer of standard hydraulic cylinders may stock 50 standard configurations (MTS) while also accepting custom orders for non-standard bore sizes, stroke lengths, or mounting configurations (MTO). The quoting challenge for the MTO portion requires the same job-similarity matching and cost estimation support as a pure job shop. The pricing challenge for the MTS portion requires the same dynamic cost modeling and competitive intelligence as a pure MTS operation.

For hybrid manufacturers, the AI tool needs to handle both workflows and know which one applies. When a customer requests a standard configuration, the system pulls from the MTS pricing model. When a customer requests a modification or custom configuration, the system switches to the MTO quoting workflow and searches for similar past custom orders. The estimator needs to see different information for each type of request, and the system must present the right interface for the right situation.

The data architecture for a hybrid tool is more complex because it draws from both the production cost database (MTS) and the custom job history (MTO), with the ability to blend information when a custom order is a modification of a standard product. The standard cylinder's production cost provides the baseline, and the custom modifications are priced incrementally based on historical data from similar customizations. This hybrid pricing approach is faster and more accurate than treating every custom order as a completely new estimating exercise.

Where Each Tool Delivers the Most Value

For MTO manufacturers, the primary value is quoting speed and accuracy. Faster quotes win more bids. More accurate quotes protect margins. The combination of speed and accuracy is worth $500,000 to $2 million per year in additional revenue for a typical MTO shop doing $8 to $15 million annually, based on the relationship between quote turnaround time and win rate.

For MTS manufacturers, the primary value is margin optimization across the product catalog. Dynamic cost modeling catches material cost increases before they erode margin. Price optimization identifies SKUs where the current price is below what the market will bear. Volume discount analysis ensures that customer pricing tiers match actual purchasing behavior. The aggregate margin improvement from these optimizations typically ranges from 1.5 to 4 percentage points across the catalog, which at $20 million in revenue represents $300,000 to $800,000 in annual margin improvement.

For hybrid manufacturers, both value streams apply. The MTO quoting tool wins more custom work at better margins. The MTS pricing tool optimizes the standard product catalog. The combined impact is typically larger than either alone because the tools reinforce each other: better cost data from the MTS side improves baseline pricing for custom modifications, and custom job data from the MTO side informs production cost estimates when standard products are modified.

Starting the Right Project

For MTO manufacturers: start with a custom AI quoting tool that connects to your ERP and quoting history. The data exists. The workflow is clear. The ROI is measurable within the first quarter. The tool replaces the research phase of quoting, not the estimator's judgment, and the speed improvement translates directly to revenue through higher win rates.

For MTS manufacturers: start with dynamic cost modeling tied to your production and purchasing data. This is the foundation for everything else. Once the cost model is in place, add competitive intelligence integration and price optimization as the second phase. The quarterly pricing review that currently takes two to three weeks can be compressed to two to three days, with better data behind every pricing decision.

For hybrid manufacturers: start with whichever side of the business generates more revenue or more quoting pain. If custom work is growing and the quoting backlog is the constraint, start with the MTO tool. If standard product margins are under pressure from material cost increases or competitive pricing, start with the MTS tool. Build the second tool after the first one is delivering value and the team is comfortable working with AI-assisted quoting and pricing.

The tools are different. The data is different. The decisions they support are different. What they share is the fundamental principle that AI in manufacturing works best when it is built around the specific way your operation runs, not around a generic model of how manufacturing works in general.