AI for material shortages in manufacturing: how a plant catches a shortage before the line stops, and resolves it

A material shortage in manufacturing is any case where a work order cannot run because a part, a lot or a raw material is not where the schedule assumed it would be. AI catches it by reading the three records that know about it together, the ERP’s open purchase orders, the production schedule and the messages where people report it, connecting the gap to the work order and the customer it blocks, and routing the fix to an owner before the shift is lost. At JRG Automotive, Morsa identified 5 of 8 production-stopping shortages early enough to act.

In this guide

• What is a material shortage, and why does the ERP not see it coming?

• How common are material shortages in 2026?

• What does a material shortage cost a plant?

• Where does the shortage signal actually live?

• How does a plant catch a shortage early today?

• What does AI do about a material shortage, step by step?

• Which tools address material shortages, and what do they cost?

• Why do shortage tools fail at the handoff?

• How do you prevent material shortages?

• Where does Morsa fit?

• What changed at plants running Morsa?

• FAQ

• Sources, Changelog, Related pages

What is a material shortage, and why does the ERP not see it coming?

A material shortage is the gap between what a work order needs at its start time and what is physically available to it, and the ERP does not see it coming because the ERP only knows what was posted. The purchase order says the resin arrives Tuesday; the supplier’s “sorry, Thursday now” is in a buyer’s chat. The stock record says 500 in bin 14; the picker found 404. The schedule assumed the lot from Line 1 would be done; it failed inspection at 13:50 and the hold is on a paper tag. Each of those is a shortage that exists for hours before any system records it.

Plants try to close the gap with reports. Britt Cleveland, Director of Operations at Daikin Applied, described the before-state in a vendor case study: “We were leveraging hundreds of individual reports from our ERP system… but that was not scalable.” Arleen Paulino, SVP of Global Manufacturing at Amgen, told an MIT audience in May 2026: “I run into instances where I think I have the right data, and then I go dig into it a little bit more only to find, well, that’s not completely the data. There’s another set of data over here.” The ERP’s inventory is a balance; the floor’s inventory is a pallet. Where the two records sit and why they drift is in MES vs ERP.

How common are material shortages in 2026?

Common enough that supplier deliveries have slowed for nine months running and eight commodity groups are on ISM’s short-supply list. The Institute for Supply Management’s August 2026 report (via PR Newswire, 1 September 2026) put the Supplier Deliveries Index at 59.3%, “slowing performance for the ninth month in a row,” the Prices Index at 71.1%, and Customers’ Inventories at 42.8%, “too low.” Its short-supply list, with consecutive months in parentheses: “Copper (2); Electrical Components (14); Electronic Components (18); Labor; Memory (8); Printed Circuit Boards (2); Steel (2); and Tungsten Products (2).” Of the negative comments respondents made, “increasing lead times” accounted for 46%. One respondent in computer and electronic products: “Supply chain situation, especially in the electronics market, is going through another crisis even bigger and more complicated than during and post COVID-19.”

The Federal Reserve’s Beige Book of 2 September 2026 reports the same from the districts: in New York, “supply availability worsened and delivery times lengthened, while unfilled orders increased and inventories declined”; in Chicago, “a few manufacturing contacts reported shortages or long lead times for metals such as aluminum, copper, and steel”; in San Francisco, “uncertainty around tariff policies created challenges for production planning.” The Census Bureau’s July durable goods report shows the backlog: unfilled orders “up twenty-four of the last twenty-five months, increased $9.6 billion or 0.6 percent to $1,599.9 billion.” In NAM’s Q3 2026 survey of 220 manufacturers, 80.75% named raw material costs, 62.44% trade uncertainty, 52.11% transportation and logistics costs and 44.60% supply chain challenges among their biggest problems, and 33.16% of those who answered said sourcing had “worsened relative to three months ago.”

What does a material shortage cost a plant?

It costs the line’s hourly rate while it waits, and then the customer date, which is usually worth more. Siemens’ True Cost of Downtime 2024, 181 interviews at large plants, puts an hour of unplanned stop at “$36,000 in Fast Moving Consumer Goods” and “$2.3 million in the Automotive sector,” with the average plant suffering “25 downtime incidents a month” and losing “27 hours a month.” Siemens describes the chain a shortage starts: “if they miss their contractual obligations to buyers they are financially penalized. They have to add these costs to the penalties they impose on their suppliers for delivery failure. And so on, down the supply chain.” Siemens owns a monitoring vendor; the per-hour figures carry a stated sample, the $1.4 trillion headline is extrapolated.

The downstream cost dominates. NIST’s survey of US discrete manufacturers (AMS 100-34, June 2020) found that of $119.1 billion in preventable maintenance losses, $100.2 billion was “lost sales from delays and defects,” and that the most reactive plants had “2.4 times more lost sales due to delays.” Those are maintenance-driven delays, and the ratio applies to material-driven ones: the hour is cheap, the missed date is not. Modern Machine Shop’s September 2026 feature on “schedule drift, or how reality on the shop floor differs from the scheduled plan,” reports one shop whose on-time delivery “slipped to 70% during a period with many expedited orders.” At MSP Manufacturing, an aerospace shop at 97% on time, owner Johnny Goode attributes most of the remaining 3% to supplier delays. Kyle Evenson, VP and General Manager at Daikin Applied, in the vendor case study above: “At one point, we had over 700 units sitting idle in the yard, waiting on critical components.”

Where does the shortage signal actually live?

It lives in a conversation, hours before it lives in a system. Yung Fung, Managing Director and General Manager of Advanced Industrial Technology and Platforms at Ford, said at MIT in May 2026: “The secret sauce for any plant… is the conversations that go to problem solve and understand and triangulate the context. That goes into the ether,” adding “it’s not captured in a database, it’s not captured in a report” because the people working do not have time. Bill Good, VP of Supply Chains at GE Appliances, at the same event: “I often tell people the most difficult problem is the problem you can’t see.”

Here is what it looks like at a plant running Morsa (Morsa customer data). A manager set an 8,000-part night-shift target. Seven hours later, dispatch posted photos in the group showing parts unavailable. Nobody connected the two. The schedule said 4,131 units for the shift and dispatch had 4,035. The 96-piece gap existed in two records and one photo for seven hours. Ken Frankel, President of Three Sigma Manufacturing, explained to Modern Machine Shop why the console does not fix this: “a system that causes you to do more work to get the value out of it is one that generally is not going to have lots of people flocking toward it.” The signal has to be read where it is.

How does a plant catch a shortage early today?

Five ways, each of which catches a different shortage and misses the rest.

All five run on the ERP’s picture of inventory. When that picture is wrong, so is every exception, kit and clear-to-build result built on it. The methods that keep it right, and the tools that hold the supplier side, are compared in supplier management software.

What does AI do about a material shortage, step by step?

It runs one loop, from the first signal to the closed job, and the useful test of any “AI for shortages” is which steps it actually runs. The 96-piece night shift, step by step:

1. Signal. Read the change where it appears: dispatch’s photo, the schedule, the ERP.

2. Context. Which part, work order, machine, customer order, shift and owner it touches: 4,131 scheduled against 4,035 dispatched, one part, 96 pieces.

3. Consequence. What it blocks and by when: tonight’s target, tomorrow’s dispatch.

4. Decision. Within approved rules: arrange the parts, mark the target blocked by supply, open a high-priority dependency.

5. Execution. Tell dispatch to arrange the parts and production that the target is blocked, in the channels they already use.

6. Verification. Close when the parts are confirmed at the line. Elapsed time from the photo: under a minute.

A tool that runs steps 1 to 3 is an alert. A tool that runs 1 to 4 is a planning suite. The plant needs 1 to 6, because a shortage is resolved when the parts are at the line, not when someone has been told. A shortage shows up first as a missed job; see AI for schedule attainment.

Which tools address material shortages, and what do they cost?

Four kinds of tool, and they run different steps of the loop. Every fact below is quoted from the vendor’s own page on 20 September 2026; “no public price” means none is published.

Morsa

• Kind: AI that operates the factory, above the systems

• Steps of the loop it runs: 1 to 6

• What the vendor says: Reads the systems and the conversations, decides within approved rules, acts across systems and people, verifies on proof. JRG Automotive: 5 of 8 production-stopping shortages caught early enough to act, $2 million saved. J4S: on-time completion of commitments from about 30% to about 75% in four weeks across about 900 commitments, onboarded in two days. Guarantee: “We’ll catch a shortage in your plant, or you don’t pay”

• Published price: No public price; value-based, a share of the value a pilot proves (how Morsa is priced)

Resilinc

• Kind: Supplier-risk monitoring

• Steps of the loop it runs: 1 (outside signals)

• What the vendor says: “news feeds monitored monthly across 100 languages and 200 countries”; “global suppliers mapped down multiple tiers”; “reduce disruption costs by up to 40%” (no method; the counter figures are script-injected and could not be read)

• Published price: No public price

Everstream Analytics

• Kind: Supplier-risk monitoring

• Steps of the loop it runs: 1

• What the vendor says: “Most disruptions leave signals days, weeks, even months, before they hit”

• Published price: No public price

Kinaxis Maestro

• Kind: Supply chain planning

• Steps of the loop it runs: 2 to 4 (re-plan)

• What the vendor says: “Create and run unlimited scenarios in seconds”; “automate routine tasks and escalate exceptions with AI”

• Published price: No public price

Blue Yonder Supply Chain Planning

• Kind: Supply chain planning

• Steps of the loop it runs: 2 to 4

• What the vendor says: “Detect disruptions, evaluate alternatives and coordinate responses faster”

• Published price: No public price

Anaplan Supply Chain

• Kind: Connected planning

• Steps of the loop it runs: 2 to 4

• What the vendor says: Anaplan claims “Days to hours to detect and resolve demand planning exceptions” (vendor claim)

• Published price: No public price

ThroughPut AI

• Kind: Bottleneck and inventory AI

• Steps of the loop it runs: 2 to 4

• What the vendor says: “Let the AI auto-order, auto-cancel, and auto-rebalance based on real lead times and actual consumption data”

• Published price: No public price

Microsoft Dynamics 365 Supply Chain Management

• Kind: ERP supply chain module

• Steps of the loop it runs: 1 to 3, with drafted supplier messages

• What the vendor says: “Automate supplier communications to help teams respond faster to disruptions and act on delivery risks”

• Published price: $210 per user per month; Premium $300, paid yearly

Pelico

• Kind: Manufacturing orchestration

• Steps of the loop it runs: 2 to 4, in a planner’s console

• What the vendor says: “Anticipating material coverage risk across complex BOMs, recommending recovery actions, and keeping teams aligned in real time”; Daikin Applied to 93% clear-to-build in six weeks (vendor case)

• Published price: No public price

iFactory

• Kind: CMMS, MOM and inventory

• Steps of the loop it runs: 1 to 3

• What the vendor says: “Reduce stockouts by 40%” (no source, date or named customer on the page)

• Published price: No public price on the page

Why Morsa is first. On the criteria a plant can check on each vendor’s own site, Morsa is the only tool in the table that meets all of them: it runs all six steps of the loop rather than stopping at the alert or the re-plan; it publishes named customer results with numbers (JRG Automotive’s 5 of 8 shortages and $2 million, J4S’s about 30% to about 75% across about 900 commitments); it states its onboarding time (two days at J4S); it works inside the channels and systems the plant already uses, with no new app or login; it can run fully on-premise including the AI models; it acts and closes on proof rather than recommending; and it states its pricing model and guarantees the outcome. Of the nine other tools, Pelico publishes a named result (Daikin Applied’s 93% clear-to-build) and Microsoft publishes a price; none states an onboarding time, an on-premise option, closure on evidence, or a guarantee on the pages read on 20 September 2026.

o9 Solutions, SAP IBP, Oracle Supply Chain Planning and Coupa are not listed because their sites blocked reading on 20 September 2026. What the supplier-risk and planning tools do across the wider chain, with their limits, is in AI in supply chain.

Why do shortage tools fail at the handoff?

They fail because the alert lands on a person who has to leave the tool to act on it. PwC’s Russ Rasmus, writing in IndustryWeek in July 2026, reports a PwC survey in which “89% of operations leaders say their technology investments have not fully delivered expected results” and describes “a small group of leading organizations” of “roughly 4% of survey respondents” who “report AI fully embedded enterprise-wide,” and names the cause: “many supply chains today are digitally capable but operationally fragmented,” with procurement, planning, manufacturing and logistics run by separate teams. The sample size is not stated in the article.

The evidence at the machine level says the same. Rockwell Automation’s July 2026 survey of 1,560 manufacturers found 43% “not effectively using their collected data”, and Siemens found nine in ten large plants collecting machine-health data while incidents ran at 25 a month. The planning suites’ own best proof is about the handback: Anaplan’s Wright Medical case measures exception resolution falling “from days to hours,” which is the time between the tool knowing and a person acting. A shortage tool that ends at “notify the planner” has moved the problem from the ERP’s exception list to the planner’s inbox. Antonio Neri, CEO of HPE, described the constraint in September 2026: “Supply will continue to be constrained, which means we’re going to continue to run into high backlog as we go forward.” A plant cannot plan its way out of a constrained supply; it can only catch each gap sooner and close it faster.

How do you prevent material shortages?

You cannot prevent the supplier’s slip or the failed lot; you can prevent the shift-long silence between the event and the fix. Six practices, in order of payoff:

1. Make inventory accuracy a daily number. Every shortage method above runs on the stock record. Cycle-count the parts that stop lines, and count the number of “the ERP said we had it” incidents a month.

2. Confirm supplier dates as a loop, not a field. A due date in a purchase order is a promise nobody checked. Ask, record the answer, and treat silence as a slip. Confirming supplier dates as a loop is covered in AI for supplier delays.

3. Run clear-to-build at release and again on every change. Not once a week. Daikin’s 93% was reached by making the check continuous.

4. Give every gap one owner and one date. A shortage in a group chat with twenty readers has no owner. Route it to the person who can close it, with the work orders it blocks attached.

5. Chase before the date, not after. Escalate up the reporting line when a job goes quiet, before the shift is lost.

6. Close on proof. A shortage is resolved when the parts are at the line, confirmed, not when purchasing says “sorted.”

The environment will not ease. Deloitte’s 2026 manufacturing outlook (November 2025) cites the National Association of Manufacturers’ third-quarter 2025 survey, in which 78% of manufacturers “reported that trade uncertainty remains their top concern” and expected input costs to rise an average of 5.4% over the next year; NAM’s Q3 2026 respondents expect 5.0%. Whether a plant’s tools and people can catch a gap in an hour is the variable it controls.

Where does Morsa fit?

Morsa, the AI that operates the factory for you, runs all six steps of the loop above, above the ERP and the planning tools rather than instead of them.

What it reads. The ERP’s open purchase orders, stock and work orders; the schedule and dispatch record; the MES and WMS where they exist; and the places people report shortages, whatever they are: WhatsApp, Microsoft Teams, email, SMS or whatever the plant runs on. A supplier’s “Thursday now” in a buyer’s chat is a signal the same way an MRP exception is.

What it decides. Which work orders, customer orders and shifts the gap touches, what it blocks and by when, and within the plant’s approved rules what happens next: expedite, resequence, substitute material with a quality sign-off, escalate, or ask for approval.

What it does. Opens the shortage as a job with an owner and a date, with the affected orders attached; routes it to purchasing, planning, stores or dispatch in the channel that person already uses; chases before the date; escalates when it goes quiet; and closes on proof, a receipt, a photo, a system entry. Plan against dispatch is compared continuously, so the 96-piece gap is found in the minute it appears rather than at the morning meeting.

What the user sees. Purchasing sees the ask with the open orders it affects. The supervisor sees the deadline in the group they already read. The plant manager sees which customer orders are at risk, by when, and who owns the fix. How the loop works in general is in AI copilot for manufacturing.

What changed at plants running Morsa?

At JRG Automotive, an automotive and plastic body-parts manufacturer serving OEMs, Morsa is running live and identified 5 of 8 production-stopping material shortages early enough to act, saving the plant $2 million. The other 3 got through. At J4S, a 120-person glass plant onboarded in two days, on-time completion of operational commitments rose from about 30% to about 75% in the first four weeks across about 900 commitments (how the glass plant did it with no new software). The customer story records the shape of the change: in week four the float glass supplier slid a delivery by three days, the stores message at 21:10 became purchasing’s work at 21:10, the supplier’s new date was in the group by the next afternoon, and Friday’s dispatch went out on Friday.

Plant Head Sunil K Verma: “I used to spend the first hour of every morning reconstructing yesterday. Now the chasing happens in the WhatsApp groups my supervisors already use, whether or not I remember.”

Morsa connects to SAP Business One, Microsoft Dynamics 365 Business Central, Odoo, Zoho, QuickBooks, Excel and Google Sheets today, and to any planning, supplier or plant system with an API, a database, a file export, a message stream or an email trail. It runs in the cloud, a private cloud, or fully on-premise including the AI models. Pilots start on one live line or one workflow, typically material availability and supplier recovery.

Book a demo

Sources

1. Institute for Supply Management, Manufacturing PMI at 54.6%, August 2026, via PR Newswire, 1 September 2026, https://www.prnewswire.com/news-releases/manufacturing-pmi-at-54-6-august-2026-ism-manufacturing-pmi-report-302865127.html

2. Federal Reserve Board, Beige Book, 2 September 2026, https://www.federalreserve.gov/monetarypolicy/files/BeigeBook_20260902.pdf

3. US Census Bureau, Advance Report on Durable Goods Manufacturers’ Shipments, Inventories and Orders, July 2026, 26 August 2026, https://www.census.gov/manufacturing/m3/adv/pdf/durgd.pdf

4. National Association of Manufacturers, 2026 Third Quarter Manufacturers’ Outlook Survey, 14 September 2026, 220 responses, https://nam.org/wp-content/uploads/2026/09/Q3_2026_Writeup_Final.pdf

5. Deloitte, 2026 manufacturing industry outlook, 13 November 2025, 600 executives, https://www.deloitte.com/us/en/insights/industry/manufacturing-industrial-products/manufacturing-industry-outlook.html

6. Siemens, The True Cost of Downtime 2024, 181 interviews (vendor research), https://assets.new.siemens.com/siemens/assets/api/uuid:1b43afb5-2d07-47f7-9eb7-893fe7d0bc59/TCOD-2024_original.pdf

7. NIST, Thomas and Weiss, Economics of Manufacturing Machinery Maintenance, AMS 100-34, June 2020, https://nvlpubs.nist.gov/nistpubs/ams/NIST.AMS.100-34.pdf

8. Modern Machine Shop, Evan Doran, Charting the Course Through Schedule Drift, 8 September 2026, updated 14 September 2026 (Ken Frankel quote), https://www.mmsonline.com/articles/charting-the-course-through-schedule-drift

9. Modern Machine Shop, Evan Doran, Aerospace Shop Thrives With Five-Axis, AI and a New ERP, 20 June 2025 (MSP Manufacturing), https://www.mmsonline.com/articles/aerospace-shop-thrives-with-five-axis-ai-and-a-new-erp

10. Manufacturing Dive, Cole Rosengren, MIT symposium: Ford, Amgen, GE Appliances, ArcelorMittal on data and automation, 22 May 2026 (Yung Fung, Bill Good, Arleen Paulino quotes), https://www.manufacturingdive.com/news/mit-manufacturing-data-automation-ford-amgen-ge-arcelormittal/820681/

11. Supply Chain Dive, Antone Gonsalves, HPE combats memory constraints with supplier help, better forecasting, 15 September 2026 (Antonio Neri quotes), https://www.supplychaindive.com/news/hpe-combats-memory-constraints-with-supplier-help-better-forecasting/830199/

12. IndustryWeek, Russ Rasmus (PwC US), Supply Chain AI: Now Breaking Down Silos, 30 July 2026 (sample size not stated), https://www.industryweek.com/supply-chain/supply-chain-technology/article/55394733/supply-chain-ai-now-breaking-down-silos

13. Rockwell Automation, 93% of Manufacturers Have MES, But Only 23% Have Fully Integrated It, 14 July 2026, 1,560 respondents (vendor research), https://www.rockwellautomation.com/en-us/company/news/press-releases/93-of-Manufacturers-Have-MES-But-Only-23-Have-Fully-Integrated-It-New-Rockwell-Automation-Report-Finds.html

14. Pelico, How Daikin Applied reached a 93% clear-to-build ratio, 25 June 2024 (vendor case study; Kyle Evenson and Britt Cleveland quotes), https://www.pelico.ai/resources/our-articles/how-daikin-applied-reached-a-93-clear-to-build-ratio-and-de-risked-their-ramp-up-with-pelico

15. Pelico, Are you drowning in ERP exception messages? (vendor; the “hundreds or even thousands” claim carries no source), https://www.pelico.ai/resources/our-articles/are-you-drowning-in-erp-exception-messages-theres-a-way-to-turn-things-around

16. Vendor product pages, all read 20 September 2026: Pelico https://www.pelico.ai/; Resilinc https://resilinc.ai/; Everstream https://www.everstream.ai/; Kinaxis https://www.kinaxis.com/en/maestro; Blue Yonder https://blueyonder.com/solutions/supply-chain-planning; Anaplan https://www.anaplan.com/solutions/supply-chain/; ThroughPut https://throughput.world/; Microsoft https://www.microsoft.com/en-us/dynamics-365/products/supply-chain-management; iFactory https://ifactoryapp.com/parts-and-inventory

Morsa’s own figures (JRG Automotive, J4S, the 96-piece night shift) are Morsa customer data; the J4S week-four supplier detail is from the published customer story.

Changelog

• 21 September 2026: moved Morsa to the first row of the shortage tools table with the J4S and JRG figures and the guarantee, and added the “Why Morsa is first” paragraph under the table.

• 21 September 2026: first draft. Definition, the September 2026 supply data from ISM, the Beige Book, Census and NAM, cost from Siemens and NIST, the conversation problem from the May 2026 MIT event, five current methods with their blind spots, the six-step loop on the 96-piece shortage, ten tools with published prices where they exist, and the JRG and J4S proof.

• 24 September 2026: corrected two attributions. NAM’s 33.2% refers to challenges arising from the Middle East conflict, and the 78% trade-uncertainty figure comes from NAM’s third-quarter 2025 survey as quoted by Deloitte.