Make?vs?print debates to ERP/PLM model with routings and AM queue; steadier schedules

Overview

An aerospace manufacturer kept debating whether parts should be made conventionally or produced through additive manufacturing without a consistent basis. Buyers compared quotes, engineering weighed manufacturability, and the additive lab tracked printer queues in isolation. Intelligex implemented a cost and lead time decision model that pulled routings and work centers from Enterprise Resource Planning (ERP), supplier lead times from sourcing records, and printer capacity and post-processing steps from the additive queue. Exceptions flowed through engineering approvals. Teams made make-versus-print calls on shared facts, downstream changes decreased, and schedules stabilized.

Client Profile

• Industry: Aerospace and defense components
• Company size (range): Multi-plant operation with centralized additive manufacturing capability
• Stage: Mature ERP and Product Lifecycle Management (PLM); growing additive footprint
• Department owner: Procurement, Supply Chain & Logistics
• Other stakeholders: Engineering and PLM, Additive manufacturing lab, Supplier quality, Program management, Finance/controlling, Quality and certification, IT applications

The Challenge

Programs needed fast decisions on whether to route parts to existing suppliers, manufacture internally with conventional processes, or use metal and polymer additive manufacturing. Each function brought a different data set. Buyers had supplier quotes and minimum order quantities; planners had standard times in routings that were not always current; the additive team had a separate view of machine time, build packing, and post-processing; engineering weighed geometry, tolerances, and material performance. Decisions dragged on, and late reversals caused rescheduling and rework.

Replacing core systems was not on the table. The ERP managed items, routings, work centers, purchasing, and cost rollups. PLM held CAD, Bills of Materials (BOMs), and changes. The additive lab scheduled builds through a standalone queue and slicer workflow. None of these systems owned a unified compare-and-commit process or ensured that constraints—qualification status, inspection plans, or certification needs—were applied consistently. The team needed a single model that translated design and demand into credible cost and lead time options and a governed way to approve exceptions.

Why It Was Happening

Data lived in silos and at different levels of detail. ERP routings and work center rates reflected batch machining and finishing but did not cover additive post-processing steps end to end. Supplier lead times were updated during sourcing events and drifted between quotes. The additive lab tracked machine hours, powder changes, heat treat, support removal, and inspection separately. Without a canonical model, each decision was a one-off analysis.

Ownership and governance were unclear. Engineering approvals for flight-critical parts, certification requirements, and inspection plans were handled through change orders, but make-versus-print decisions were made in meetings and email threads. The lack of a formal review gate meant exceptions were discovered late, and cost rollups in ERP diverged from the reality of how parts were built. Debate focused on defending inputs rather than choosing the best option.

The Solution

Intelligex deployed a decision support layer that assembled cost and lead time scenarios for each part and demand window, using data from ERP, PLM, supplier records, and the additive lab. The model calculated manufacturing and purchase options side by side, including routings, setup and run times, queue and changeover effects, additive build packing, and all post-processing steps. Engineering approvals were required when parts had critical characteristics, new materials, or unqualified additive routes. The approach introduced a consistent, auditable choice without replacing existing tools.

• Integrations: Bi-directional sync with ERP for items, routings, work centers, cost elements, open demand, and purchase info records (for example, SAP S/4HANA); PLM for geometry, material, tolerances, and change effectivity (for example, PTC Windchill); additive lab queue and post-processing steps for machine availability, build parameters, and downstream operations; supplier lead times and minimum order quantities from sourcing records.
• Cost and lead time model: Canonical structure covering setup/run times, toolpath or build time, changeovers, post-processing (heat treat, support removal, machining, surface finish), inspection, outside processing, scrap risk, and logistics.
• Qualification and constraints: Flags for part criticality, allowable materials and processes, inspection plans, and certification paths aligned with additive manufacturing guidance from the NIST Additive Manufacturing program; guardrails prevented unqualified routes without approval.
• Scenario workflows: Side-by-side options for internal make, external purchase, and additive, each with earliest available dates and cost breakdown; sensitivity to demand window and lot size.
• Validations and guardrails: Checks for missing routings, outdated supplier lead times, unmodeled post-processing steps, and incompatible materials; prompts to update ERP or supplier records when gaps were found.
• Review gates: Human-in-the-loop approvals for flight-critical parts, first-time additive builds, and deviations from standard inspection or certification plans; reason codes captured.
• Dashboards: Visibility into decision status by program, parts routed to additive versus conventional paths, qualification progress, and exceptions aging.
• Permissions and audit: Role-based access for engineering, procurement, planners, and the additive lab; immutable logs linking inputs, decisions, and approvals back to ERP and PLM references.

Implementation

• Discovery: Mapped the decision flow from demand signal through final route selection; inventoried ERP routings and work centers for target part families; documented additive build steps and post-processing; collected supplier lead time data and quote practices; identified qualification and certification checkpoints.
• Design: Defined the canonical cost and lead time model, including post-processing and inspection; designed data mappings from ERP, PLM, and the additive queue; set approval criteria for critical characteristics and first-time routes; established a shared glossary for statuses, reason codes, and constraints.
• Build: Implemented connectors to ERP, PLM, and the additive lab schedule; configured the model for machining, forming, and additive processes; built the scenario and approval workflows; created dashboards and audit logging.
• Testing/QA: Replayed past make-versus-print decisions and validated model outputs against actual outcomes; ran observe-only recommendations while teams continued legacy analysis; executed engineering reviews for first-time additive routes to tune guardrails.
• Rollout: Piloted with selected part families and programs; kept existing spreadsheets and meetings as a fallback; enabled gating for flight-critical and first-time additive decisions after results matched expectations; expanded to additional parts and sites as exceptions stabilized.
• Training/hand-off: Scenario-based sessions for buyers, planners, engineers, and the additive team; quick guides explaining model assumptions and approval triggers; runbooks for updating routings, supplier lead times, and qualification status; transitioned operations to supply chain and engineering with IT support on call.

Results

Teams made decisions with a single view of costs, lead times, and constraints. The model exposed the real drivers—post-processing queues, supplier minimums, inspection bottlenecks—so discussions focused on trade-offs rather than reconciling spreadsheets. Engineering reviews were triggered only when required by part criticality or process novelty, and approvals carried a clear rationale.

Late changes and rescheduling declined. The route chosen at gate review held through execution because inputs matched reality: supplier lead times were current, routings included all steps, and the additive queue and post-processing were part of the plan. Program managers gained predictable dates, procurement avoided rework on purchase orders, and the additive lab ran a steadier schedule.

What Changed for the Team

• Before: Debates hinged on disconnected spreadsheets and emails; After: A shared model produced side-by-side options with traceable inputs.
• Before: Additive steps and post-processing were estimated separately; After: Build and post-processing were modeled end to end alongside conventional routes.
• Before: Supplier lead times drifted between quotes; After: Lead times and minimums flowed from sourcing records into every scenario.
• Before: Engineering approvals were late or blanket; After: Approvals were triggered by clear criteria and captured with reason codes.
• Before: ERP routings lagged behind real operations; After: Validations flagged missing steps and prompted updates before decisions were made.

Key Takeaways

• Compare internal make, external purchase, and additive paths in a single model that includes post-processing and inspection, not just machine time.
• Integrate ERP routings, supplier lead times, and additive queue data; do not replace existing systems.
• Gate decisions with engineering approvals tied to part criticality and first-time routes, and capture reasons for traceability.
• Use validations to surface stale routings and lead times and push corrections upstream before committing.
• Pilot on a few part families, run recommendations in observe-only mode, then enforce review gates once behavior matches expectations.

FAQ

What tools did this integrate with?
The decision layer connected to ERP for items, routings, work centers, cost elements, and sourcing data (for example, SAP S/4HANA), to PLM for geometry, material, tolerances, and change effectivity (for example, PTC Windchill), and to the additive lab’s scheduling and post-processing tracker. Supplier lead times and minimums flowed from purchasing records or the SRM. The model referenced additive process concepts consistent with the NIST Additive Manufacturing program.

How did you handle quality control and governance?
We established engineering approval gates for flight-critical parts, first-time additive routes, and deviations from inspection or certification plans. Validations checked for missing routing steps, stale lead times, and unqualified processes. Every decision and exception carried a reason code and was audit-logged, with references back to ERP routings, PLM revisions, and supplier records.

How did you roll this out without disruption?
We ran the model in observe-only mode first, generating recommendations alongside the legacy analysis. Teams compared outputs on real parts and tuned assumptions. Once comfortable, we enabled gating for defined categories while keeping the fallback path available. Coverage expanded as results proved consistent and exceptions declined.

How did the model account for additive specifics like packing and post-processing?
The additive pathway included build time based on part geometry and build parameters, packing effects across jobs, changeovers for material and machine, and the full chain of post-processing steps such as heat treat, support removal, finish machining, and inspection. Those steps carried capacity and queue assumptions from the additive lab and outside processors.

What about parts with certification or qualification requirements?
Parts flagged as critical or subject to specific certification paths triggered engineering and quality review. Unqualified additive routes were blocked without approval, and the model surfaced the required inspection plans and allowable materials. Decisions captured the rationale and any interim controls, creating a traceable record for audits and program reviews.