EngineeringPilot Pilot, prototype, early rollout or time-limited trial; results are limited or not yet measured.

AI traces each engineering change to the parts, purchase orders, routings and work instructions it affects, before it reaches the floor (custom-assembly manufacturer, Cytovale)

Added 2026-10-23 · Last updated 2026-10-23

Manufacturers in this case: Cytovale · Galvanize Therapeutics · all manufacturers using AI

Role

Engineering (engineering change management: ECR/ECO impact analysis) · Typical users: design engineer, configuration or change manager, manufacturing engineer, buyer or planner, quality or regulatory engineer, engineering manager

The problem

An engineering change rarely stays in engineering. A last-minute part change can delay a purchase order, a moved sub-assembly can invalidate the work instructions, and if nobody connects the change to the open orders, the shop floor discovers it after the job has started, with rework and scrap. Finding every affected item means digging through the BOM, where-used lists, routings, open purchase and production orders, requirements and test documents, often in different systems. A mid-sized manufacturer of custom, complex assemblies was changing designs days before builds and had no quick way to see the downstream impact. At Cytovale, a medical diagnostics device maker, working out which requirements, risks and tests a proposed change affects took up to weeks, depending on the change, and required several subject-matter experts.

How AI helps

The AI connects the change to the product data around it. At the custom-assembly manufacturer, a dependency graph built from change logs, BOMs, procurement lead times and vendor data, routings, work instructions and production orders is combined with text analysis of the design notes and business rules: every new change is scanned automatically, and buyers, planners and production get an alert within minutes for the high-risk ones. At Cytovale, an AI agent reads hardware descriptions, software specifications, requirements and risk documents linked in one traceability tool, and drafts the impact assessment: affected requirements, a proposed test strategy and the rationale, which engineers then review. Engineers and the change board still approve every change.

Step by step

1. Pick the changes that hurt most: look back at 6 to 12 months of ECOs and list the ones that caused late purchases, wrong builds, rework or scrap, and why the impact was missed.
2. Make the links explicit: where-used from the BOM, open purchase and production orders by part, routings and work instructions by operation, and, in regulated products, requirements, risks and tests.
3. Start with a checklist report: for each new change, the AI lists every affected item and open order and drafts the impact summary for the change owner.
4. Test it on past changes first, as the custom-assembly manufacturer did, and count which known problems it would have caught.
5. Define who gets alerted for which kind of impact (buyer, planner, production, quality) and set thresholds so low-risk changes do not flood people.
6. Keep the change board's approval as it is, and measure missed impacts, rework from changes and change cycle time against your baseline.

Tools

• Captivix custom change-impact model (dependency graph, NLP on design notes, business rules) connected to ERP and PLM: used at the custom-assembly manufacturer
• Ketryx (traceability and AI agents for regulated product development, with Polarion): piloted at Cytovale
• Arena PLM API with an in-house item-check app: used at Galvanize Therapeutics
• Other tools to evaluate: AI assistants in your PLM, for example Siemens Teamcenter Copilot for BOM (impact analysis and part replacement with human confirmation), and the change modules of PTC Windchill, Aras or Arena
• Any LLM (ChatGPT, Claude, Copilot, Gemini) with the prompt below, on one change request and exported where-used and open-order lists, for a first test

In the AI tools index: Arena PLM · Captivix · all AI tools used by manufacturers

As reported by the source; most figures are vendor or self-reported and not independently verified.

Human review: cautions

• The AI drafts the impact analysis; the change owner and the change board still check it and approve every change.
• An impact the data does not show cannot be found: missing where-used links, work instructions outside the PLM or informal shop-floor practices will be missed.
• In regulated products, the impact assessment is a controlled record: validate the tool and review every AI output, as Cytovale does.
• The custom-assembly manufacturer's figures are published by the integrator that built the system; the Cytovale figures come from a vendor case study of a pilot; the Galvanize app is not described as AI. Do not paste confidential drawings or specifications into public AI tools.

You are an experienced engineering change coordinator at a manufacturer.

I will paste: (1) one engineering change request (what changes, why, old and new part numbers and revisions, proposed effectivity); (2) the where-used list of the affected parts; (3) open purchase orders and open production orders for those parts and their parents [order, part, quantity, due date, status]; (4) the routings and work instruction numbers that mention them; (5) on-hand stock of the old parts.

Do this:
1. List every affected item: parent assemblies, other products using the part, routings and work instructions to update, quoting the line from my data for each.
2. List every open purchase and production order affected, and for each say whether it is likely already in progress, based on its status and date.
3. Show the stock of old parts that could become obsolete; copy quantities as given and give me a spreadsheet formula for any total.
4. Propose effectivity options (use up old stock, change at a date, change immediately) with the consequences of each.
5. Draft a short impact summary for the change board and a notification for each team (purchasing, planning, production, quality).
6. List what I did not give you that could hide an impact, as questions.

Never approve or release the change yourself.
Use only the data I paste in; if something is missing, write 'not found' or list it as a question instead of guessing. Do not send, order, post or change anything in any system: produce drafts for my review and approval.

Related cases

• AI pulls BOM data from supplier documents and checks it against stock and specs
• AI runs a first-pass check of drawings and models before the design review, so engineers discuss real issues (INFICON, McDermott)
• AI-drafted SOPs and work instructions from existing documents
• Engineers ask questions of their standards, drawings and past project files and get an answer with the source page (Flender, NETZSCH, Victaulic)

All 11 Engineering use cases →

Sources

• https://captivix.com/industries/manufacturing/use-cases/engineering-change-impact
• https://www.ketryx.com/case-studies/how-cytovale-transformed-change-impact-assessment-ketryx-ai
• https://www.ptc.com/en/case-studies/galvanize-harnesses-cloud-plm-and-ai
• https://blogs.sw.siemens.com/teamcenter/ai-bom-agent-plm-engineering-change/