# Worked Industrial AI acceptance pack: fictional unplanned-stop triage on one packaging line

**FICTIONAL, ILLUSTRATIVE EXAMPLE. No client, no real company, no real plant or fleet and no real benchmark. Every figure is an invented input or is calculated from invented inputs.**

Prepared 2026-10-10; human review not recorded. Prepared with AI assistance. It is not a client deliverable, a customer result or an approval to operate, buy or sell anything.

For: Maintenance manager as operating owner, plant manager as sponsor, and the line's technicians

Decision: Should a triage assistant for unplanned stops on one packaging line go into daily use, should the plant keep the process fix alone, or should the work stop?

Gate: productization; mandate: product-leadership-program.

Number format: the comma separates thousands and the point marks decimals; 4,000 EUR means four thousand euros.

## 1. The decision and who owns it

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

A fictional mid-sized plant that packs consumer goods. One cartoning line runs two shifts, five days a week. The workflow is the triage of unplanned stops longer than five minutes: from the moment the line stops to the moment a technician has confirmed the cause.

- **Maintenance manager.** Operating owner. Sets acceptance with the technicians, takes the decision and runs whatever is kept.
- **Plant manager.** Sponsor. Agrees the value of a downtime hour and the two economic thresholds before the trial.
- **Line maintenance technicians.** Users. Take part in the trial, record each stop and say when a suggestion was wrong.
- **IT and operational-technology security lead.** Approves what the assistant may read, where it runs and who can open it.
- **Safety officer and quality manager.** Keep their authority. Either can suspend the trial without giving a business reason.
- **Integrator.** Installs and connects the candidate under the plant's purchase order.

## 2. The baseline, before anything changes

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

Invented measurements over 8 weeks, with the way each would be collected in a real plant.

| Measure | Invented value | How it would be collected |
| --- | --- | --- |
| Unplanned stops longer than five minutes | 208 in 80 shifts (2.6 a shift) | The line controller's stop counter, reconciled each shift with the technician's log so that no stop is dropped. |
| Mean time from stop to confirmed cause | 14 minutes | Two timestamps per stop on the line terminal: line stopped, cause confirmed. |
| Mean time from confirmed cause to restart | 11 minutes | The same log, with the restart time taken from the line controller. |
| Downtime in the window | 86.67 hours, 13.54% of 640 scheduled hours | Calculated from the three rows above. See the check table. |
| Stops resolved by the first corrective action | 141 of 208 (67.79%) | The technician's closing entry: did the first action restart the line and keep it running for the rest of the shift? |
| Stops where the technician searched manuals or the shift log | 96 of 208 (46.15%) | A tick box on the stop record, checked against one week of direct observation. |
| Stops from the five most frequent causes | 131 of 208 (62.98%) | Cause codes assigned by the maintenance engineer from the closing entries. |
| Value of one hour of line downtime | 900 EUR | Set by the plant's finance controller from the line's contribution. It is the plant's own figure and is agreed before the trial. |

## 3. Non-AI comparators, considered first

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

### Fix the process

Standard cause codes on the line terminal and a one-page fault-finding card for each of the five most frequent causes, written by the maintenance engineer with the technicians.

The five causes cover 62.98% of stops; the other 77 stops in the window have no dedicated card. This is a coverage limit, not by itself a reason to buy AI.

### Use a tool the plant already has

The text search of the maintenance system the plant already licenses, opened on a tablet at the line, over past work orders.

This search reaches less frequent causes, but past work orders are written inconsistently and it does not search the machine suppliers' manuals.

### Both together, measured

The two were run together for 4 weeks before any new technology was selected. It cost the plant 10 staff days (4,000 EUR at the internal day cost).

Mean diagnosis time fell from 14 to 10.5 minutes over 104 stops, and first fixes that held rose to 73.08%. On the annual stop count that is worth 62,790 EUR a year. This is the result the candidate has to beat, not the original baseline.

## 4. The candidate capability

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

A triage assistant on the line terminal. The technician enters the stop code and a short description; the assistant searches the suppliers' manuals, past work orders and the shift log, and returns up to three likely causes, each with the page or record it comes from and the plant's own check procedure for that cause.

- It reads documents. It is not connected to the line controller and cannot start, stop or change anything.
- A suggestion is never a command. The technician decides what to check and what to do.
- When it finds no supporting document it says so and suggests nothing.

## 5. Acceptance measures, thresholds and evidence

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

The thresholds were agreed with the operating owner before the trial. The observed values are invented.

In this fictional trial, each of the 71 consultations produces one response: 66 contain up to three proposed causes and 5 abstain. The top-three rate is counted per non-abstaining response, and 49 of them contain the confirmed cause. The citation check and the review for contradictions with procedures cover every proposed cause in every non-abstaining response.

| Measure | Threshold | Observed (invented) | Met | Evidence needed | Evidence owner |
| --- | --- | --- | --- | --- | --- |
| Mean time from line stop to confirmed cause, with the assistant | At most 9 minutes | 8.9 minutes over 100 stops | Yes | Stop log with a timestamp at line stop and at cause confirmation, entered on the line terminal. Every stop over five minutes counts; none is excluded afterwards. | Maintenance manager |
| Minutes gained per stop over the non-AI comparator | At least 1 minute | 1.6 minutes (10.5 against 8.9) | Yes | The same log for the comparator weeks: same line, same shift pattern, same definition of a stop and of a confirmed cause. | Maintenance manager |
| Stops resolved by the first corrective action | At least 75%, and not below the comparator | 78% (comparator 73.08%) | Yes | The technician's closing entry for each stop, checked weekly by the maintenance engineer. | Maintenance engineer |
| Responses whose up to three proposed causes include the confirmed cause | At least 70% of non-abstaining responses | 74.24% (49 of 66) | Yes | Each response is stored with its sources and matched to the confirmed cause by the maintenance engineer, who did not write the response. | Maintenance engineer |
| Responses in which every proposed cause cites a plant document the technician can open | 100% of non-abstaining responses; with no supporting document the assistant says so and proposes nothing | 100% (66 of 66), with 5 consultations where it declined to propose a cause | Yes | An automatic check of every proposed cause in every stored response for a source reference that opens the cited page. It shows that the reference opens, not that the page supports the cause. | IT and operational-technology security lead |
| Proposed causes that contradict a safety or lockout procedure | 0 | 0 contradictions across every proposed cause in 66 non-abstaining responses | Yes | The safety officer reads every proposed cause in every response issued in the trial, not a sample, against the plant's procedures. The review records contradictions with those procedures; it is not a safety assessment of the assistant. | Safety officer |
| Availability during scheduled shifts, with the fallback exercised | At least 98% of scheduled minutes, and the fallback used at least 2 times | 98.75% (240 of 19,200 minutes unavailable); fallback used 3 times | Yes | The service log of unavailable minutes kept by IT, and a stop record for each time the fallback was used. | IT and operational-technology security lead |
| Stops where the technician consulted the assistant | At least 60% | 71% (71 of 100) | Yes | The assistant's query log matched to the stop log, and a short interview with each technician on the line. | Maintenance manager |
| Annual value beyond the comparator, against the plant's annual running cost | At least 1.5 times the running cost | 1.37 times (28,704 EUR against 21,000 EUR) | No | The check table below, with the value of a downtime hour confirmed by the plant's finance controller and the cost lines confirmed by purchasing. | Plant manager |
| Years for the net annual value to repay the plant's one-off cost | At most 2 years | 3.46 years (26,650 EUR against 7,704 EUR a year) | No | The same check table, reviewed by the finance controller. | Plant manager |

The automated citation check establishes that references open. It does not establish source relevance, currency, support for each claim or correct abstention; those require additional evaluation before operational use.

## 6. Operator procedure

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

### At every stop

1. Make the machine safe with the plant's existing procedure. The assistant plays no part in this step and is not opened until it is done.
2. Enter the stop code and a one-line description on the line terminal.
3. Read the suggestions and open the cited source for the one you intend to act on.
4. Decide. You may ignore every suggestion.
5. Carry out the check and the repair under the plant's normal rules.
6. Close the stop record: the confirmed cause, whether the first action held, and whether the suggestion was right, wrong or not used.

### When the assistant is wrong

1. Mark the suggestion wrong on the stop record. This takes one tap and is never held against the technician.
2. Carry on with the fault-finding card or the maintenance system search, as before.
3. Call the maintenance engineer under the existing escalation rule. The assistant does not change who is called or when.
4. The maintenance engineer reviews every suggestion marked wrong each week, distinguishing source errors from retrieval, interpretation and configuration errors. The document owner approves any source correction; valid source documents are retained. The integrator corrects assistant defects, and affected examples are retested before the change is accepted. Safety contradictions trigger immediate suspension under the stop conditions.

### When the assistant is unavailable

1. Do not wait for it. Use the fault-finding cards and the maintenance system search, which stay at the line for this reason.
2. Tick the unavailable box on the stop record so the minutes are counted.
3. IT restores the service outside the stop. No stop is extended to bring the assistant back.

## 7. Data, access and safety boundaries

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

### Data

- Sources are the suppliers' manuals the plant is licensed to hold, the plant's work orders and the shift log. Nothing else is indexed.
- The index and the query log stay inside the plant network. Whether any text may leave it is decided by the IT and operational-technology security lead, in writing, before the trial.
- Shift logs and work orders contain technicians' names. The plant's data-protection contact decides what is removed or masked before indexing and how long query logs are kept, and the technicians are told what is recorded.

### Access

- Read-only accounts. The assistant cannot write to the maintenance system or the shift log.
- No connection to the line controller or to any safety function.
- Only the line's technicians, the maintenance engineer and the maintenance manager can open it.

### Safety and quality

- Lockout and the plant's safety procedures are unchanged and come first.
- The safety officer reads every suggestion issued in the trial and can suspend it at once.
- Quality release decisions stay with the quality manager. The assistant gives no opinion on product release.

## 8. Cost lines the plant carries

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

These are costs the plant itself carries. No consulting fee and no Hyperion price is part of this pack: a fee for leading such work is quoted separately in writing and would be added by the plant as its own line.

| One-off line | Basis (invented) | EUR |
| --- | --- | ---: |
| Integrator work | 15 days at 950 EUR | 14,250 |
| Maintenance engineer | 12 days at 400 EUR | 4,800 |
| Technicians in the trial | 6 days at 400 EUR | 2,400 |
| IT and operational-technology security | 5 days at 400 EUR | 2,000 |
| Cleaning up the source documents | 8 days at 400 EUR | 3,200 |
| **One-off total** | | **26,650** |

| Annual line | Basis (invented) | EUR a year |
| --- | --- | ---: |
| Software licence | Supplier quote, invented for this example | 9,600 |
| Hosting on the plant's own server | IT estimate, invented for this example | 1,800 |
| Keeping the documents current | 2 days a month at 400 EUR | 9,600 |
| **Annual total** | | **21,000** |

The payback of 3.46 years in the measures above is a total-project calculation. At the post-trial decision, distinguish expenditure already incurred from avoidable future costs; do not treat spent trial effort as a new purchase cost. The comparator's 62,790 EUR is gross modelled annual benefit before any ongoing upkeep. A real comparison must identify recurring costs for both options and separate shared costs from incremental assistant costs.

## 9. Stop conditions during the trial

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

- One suggestion that contradicts a safety or lockout procedure: the trial is suspended the same shift and does not restart without the safety officer's written agreement.
- Text from the plant's documents or logs found outside the boundary the security lead approved: the assistant is switched off and the index removed.
- First fixes that hold fall below the comparator for two weeks running: the trial stops and the line returns to the comparator procedure.
- Availability below 98% of scheduled minutes over the trial: no decision to go is possible, whatever the other measures say.
- The comparator and the candidate both fail to improve on the baseline: the work stops and the plant keeps its original procedure.

## 10. The worked decision

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

The rule, agreed before the trial:

- Go: every acceptance and economic measure is met. The assistant goes into daily use on this line with the procedure above.
- Fix the process: the comparator beat the baseline and the assistant did not clear every measure. The plant keeps the cause codes, the cards and the search, and does not buy the assistant now.
- Stop: a safety contradiction occurred, or the assistant did not clear every measure and the comparator did not beat the baseline either.

**Outcome in this example: Fix the process.**

In this fictional example, all eight listed acceptance measures meet their thresholds: mean diagnosis time is 1.6 minutes shorter than with the comparator, technicians consulted the assistant at 71 of 100 stops, and the review recorded no contradiction with the plant's safety or lockout procedures. This does not establish that the assistant or its suggestions are safe. A safety assessment by the plant's safety officer would still have to establish that acting on a proposed cause cannot lead a technician to skip or shorten isolation and lockout, that a wrong or outdated proposal is caught before anyone acts on it, and that the procedures used for the review are complete and current. The two economic measures are not met. Beyond the process fix it is worth 28,704 EUR a year against a running cost of 21,000 EUR, a ratio of 1.37 where 1.5 was agreed, and the one-off cost of 26,650 EUR would take 3.46 years to repay where 2 was agreed. The process fix alone is worth 62,790 EUR a year for 4,000 EUR of staff time. The maintenance manager's decision is to fix the process: keep the comparator procedure on this line and not put the assistant into use now.

The following changes trigger a fresh review, not approval. The first two address only the running-cost ratio; the payback gate of at most 2 years must also pass. With all stated costs unchanged, both economic gates require at least 34,325 EUR of annual value beyond the comparator: a gain of at least 1.92 minutes per stop at 1,196 stops a year, or at least 1,431 stops a year at a gain of 1.6 minutes. Expansion costs must be included before applying these thresholds.

- The gain over the comparator reaches 1.76 minutes a stop at the present stop count, approximately the threshold for the value to reach 1.5 times the running cost.
- The same licence and upkeep serve at least 1,313 comparable stops a year at the observed gain, for example because further lines with the same machines are added. That would have to be measured on those lines, not assumed.
- The running cost changes, or the plant's finance controller revises the value of a downtime hour.

## 11. Check table

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

Every derived figure above, with its formula. Results are rounded to two decimals.

| Check | Formula | With this example's numbers | Result |
| --- | --- | --- | ---: |
| shifts-observed | weeks × days per week × shifts per day | 8 × 5 × 2 | 80 shifts |
| scheduled-hours | shifts observed × hours per shift | 80 × 8 | 640 hours |
| stops-per-shift | baseline stops ÷ shifts observed | 208 ÷ 80 | 2.6 stops |
| baseline-downtime | baseline stops × (mean diagnosis + mean repair minutes) ÷ 60 | 208 × (14 + 11) ÷ 60 | 86.67 hours |
| downtime-share | downtime hours ÷ (shifts observed × hours per shift) × 100 | 86.67 ÷ (80 × 8) × 100 | 13.54 % |
| baseline-first-fix | first fixes that held ÷ baseline stops × 100 | 141 ÷ 208 × 100 | 67.79 % |
| document-search-share | stops with a document search ÷ baseline stops × 100 | 96 ÷ 208 × 100 | 46.15 % |
| top-five-cause-share | stops from the five most frequent causes ÷ baseline stops × 100 | 131 ÷ 208 × 100 | 62.98 % |
| stops-without-a-card | baseline stops − stops from the five most frequent causes | 208 − 131 | 77 stops |
| stops-per-year | stops per shift × production weeks × days per week × shifts per day | 2.6 × 46 × 5 × 2 | 1,196 stops |
| comparator-gain | baseline − comparator diagnosis minutes | 14 − 10.5 | 3.5 minutes per stop |
| comparator-first-fix | comparator first fixes that held ÷ comparator stops × 100 | 76 ÷ 104 × 100 | 73.08 % |
| comparator-staff-cost | comparator staff days × internal day cost | 10 × 400 | 4,000 EUR |
| comparator-annual-value | stops per year × (baseline − comparator diagnosis minutes) ÷ 60 × value of a downtime hour | 1,196 × (14 − 10.5) ÷ 60 × 900 | 62,790 EUR a year |
| candidate-gain | comparator − candidate diagnosis minutes | 10.5 − 8.9 | 1.6 minutes per stop |
| candidate-first-fix | candidate first fixes that held ÷ candidate stops × 100 | 78 ÷ 100 × 100 | 78 % |
| cause-in-top-three-share | responses containing the confirmed cause ÷ non-abstaining responses × 100 | 49 ÷ 66 × 100 | 74.24 % |
| cited-source-share | responses with a source for every proposed cause ÷ non-abstaining responses × 100 | 66 ÷ 66 × 100 | 100 % |
| consultation-share | stops where the assistant was consulted ÷ candidate stops × 100 | 71 ÷ 100 × 100 | 71 % |
| candidate-annual-value | stops per year × (comparator − candidate diagnosis minutes) ÷ 60 × value of a downtime hour | 1,196 × (10.5 − 8.9) ÷ 60 × 900 | 28,704 EUR a year |
| candidate-scheduled-minutes | candidate weeks × days per week × shifts per day × hours per shift × 60 | 4 × 5 × 2 × 8 × 60 | 19,200 minutes |
| availability | (1 − unavailable minutes ÷ (weeks × days × shifts × hours × 60)) × 100 | (1 − 240 ÷ 19,200) × 100 | 98.75 % |
| integrator-cost | integrator days × integrator day rate | 15 × 950 | 14,250 EUR |
| maintenance-engineer-cost | maintenance engineer days × internal day cost | 12 × 400 | 4,800 EUR |
| technicians-cost | technician days × internal day cost | 6 × 400 | 2,400 EUR |
| it-and-ot-security-cost | IT and operational-technology security days × internal day cost | 5 × 400 | 2,000 EUR |
| document-clean-up-cost | document clean-up days × internal day cost | 8 × 400 | 3,200 EUR |
| plant-staff-days | maintenance engineer + technician + IT and security + document clean-up days | 12 + 6 + 5 + 8 | 31 days |
| one-off-cost | integrator days × integrator day rate + plant staff days × internal day cost | 15 × 950 + 31 × 400 | 26,650 EUR |
| annual-upkeep-cost | upkeep days per month × 12 × internal day cost | 2 × 12 × 400 | 9,600 EUR a year |
| annual-running-cost | licence + hosting + upkeep days per month × 12 × internal day cost | 9,600 + 1,800 + 2 × 12 × 400 | 21,000 EUR a year |
| net-annual-value | candidate annual value − annual running cost | 28,704 − 21,000 | 7,704 EUR a year |
| value-to-running-cost | candidate annual value ÷ annual running cost | 28,704 ÷ 21,000 | 1.37 times |
| payback | one-off cost ÷ (candidate annual value − annual running cost) | 26,650 ÷ (28,704 − 21,000) | 3.46 years |
| gain-needed-to-reopen | required ratio × annual running cost ÷ value of a downtime hour × 60 ÷ stops per year | 1.5 × 21,000 ÷ 900 × 60 ÷ 1,196 | 1.76 minutes per stop |
| stops-needed-to-reopen | required ratio × annual running cost ÷ value of a downtime hour × 60 ÷ observed gain, rounded up | 1.5 × 21,000 ÷ 900 × 60 ÷ 1.6 | 1,313 stops a year |
| value-needed-for-both-gates | max(required ratio × annual running cost; annual running cost + one-off cost ÷ maximum payback years) | max(1.5 × 21,000; 21,000 + 26,650 ÷ 2) | 34,325 EUR a year |
| gain-needed-for-both-gates | value needed for both gates ÷ value of a downtime hour × 60 ÷ stops per year, rounded up to two decimals | 34,325 ÷ 900 × 60 ÷ 1,196 | 1.92 minutes per stop |
| stops-needed-for-both-gates | value needed for both gates ÷ value of a downtime hour × 60 ÷ observed gain, rounded up | 34,325 ÷ 900 × 60 ÷ 1.6 | 1,431 stops a year |

## 12. What this pack does not prove

**Fictional worked example: invented inputs and calculated illustrations; no customer evidence.**

- Nothing here was measured. The plant, the line, the stops, the times, the costs and the trial results are invented to show the shape of an acceptance record.
- The three periods run one after the other, so the stop mix differs between them. A real comparison has to show that the periods are comparable, or alternate the procedures by shift.
- The comparison rests on means over about 100 stops. A real review needs the spread of the times, not only the mean, before treating a difference of 1.6 minutes as real.
- A minute of diagnosis saved is counted as a minute of line running regained. The calculation assumes diagnosis is on the critical path to restart and that earlier diagnosis brings restart forward by the same amount, without increasing repair time, repeat faults or quality losses. The plant must verify that relationship using total stop-to-restart time and establish whether the recovered capacity creates the assumed contribution.
- The illustrative first-fix measure uses the remainder of the shift as follow-up, so exposure differs between stops. A real trial should agree a common follow-up period in operating hours, distinguish recurrence of the same fault from unrelated faults, and identify incomplete follow-up before comparing first-fix rates.
- The candidate combines an assistant with document clean-up and broader source access. The example does not isolate the contribution of AI; a real comparison should also consider making the same cleaned manuals and records searchable without an assistant.
- The annual figures apply the baseline stop rate to a full year. They are not a forecast of stops, and they say nothing about a different line, plant or product.
- The thresholds are the fictional plant's choices. A real operating owner sets different ones for a different workflow and risk.
- This pack gives no legal, data-protection, safety or conformity opinion. Those decisions stay with the people named for them.

Payload SHA-256: f63794bb10931e8191d85e3ffc1ceddd0c6c835bb0fc56a1b35e3720d1f14d5e