# AI scores ergonomic risk from a phone video of the job, so safety teams fix the worst lifts before they cause sprains and strains (Superior Tube Products, Latham Pools, Johns Manville)

> AI scores ergonomic risk from a phone video: Superior Tube cut workers' comp costs 98%, Latham Pools sprains and strains 91%, Johns Manville injury risk 30%.

- Canonical URL: https://aiagentsformanufacturing.com/use-cases/ai-video-ergonomic-risk-assessment-sprains-strains-superior-tube-latham-johns-manville/
- Department: Production (https://aiagentsformanufacturing.com/departments/production/)
- Evidence level: Verified. Real deployment at a named or precisely described company, with a measured result from a primary source.
- Added: 2026-11-17
- Last updated: 2026-11-17
- Publisher: AI Agents for Manufacturing (https://aiagentsformanufacturing.com)
- Manufacturers named: Superior Tube Products (https://aiagentsformanufacturing.com/companies/#superior-tube-products), Latham Pools (https://aiagentsformanufacturing.com/companies/#latham-pools), Johns Manville (https://aiagentsformanufacturing.com/companies/#johns-manville)
- AI tools deployed: TuMeke (https://aiagentsformanufacturing.com/tools/#tumeke), Humantech motion capture (https://aiagentsformanufacturing.com/tools/#humantech-motion-capture)

## Role

Health and safety (EHS) and production supervision (ergonomics, sprains and strains, workers' compensation). Typical users: EHS manager, health and safety coordinator, plant manager, production supervisor, industrial engineer, HR or workers' compensation manager

## The problem

Sprains, strains and other musculoskeletal disorders come from lifting, bending, reaching and repeating the same motion all shift, and they are often the most common and costly injuries in a plant. The classic assessments (the revised NIOSH lifting equation, REBA) need a trained person, detailed observation and calculations for each task, which small plants rarely have time or staff for. At Superior Tube Products (Davenport, Iowa, USA), an employee-owned maker of bent and laser-cut tubes with a team of over 50, employees lift steel tubing and parts weighing up to 50 pounds individually, hold static postures at heavy machines and repeat motions when loading parts, doing fit checks or using hand tools; new hires are most at risk. At Latham Pools, a maker of in-ground swimming pools with more than 20 locations, sprains and strains were the highest share of injuries across the organization, none of the locations had an ergonomist, and there was no video or data to check what caused a reported injury. At Johns Manville, the EHS team's manual analysis took roughly 45 minutes per recorded task; the company cites the National Safety Council that overexertion is the leading cause of injury.

## How AI helps

Someone records a short video of a worker doing the task with a phone or tablet. Computer vision estimates the body's posture frame by frame, and the software scores the risk by body part (back, shoulders, wrists, for example) using standard ergonomic methods, flags the motions that drive the score and helps rank tasks for correction. At Superior Tube Products, short videos of high-risk tasks (loading and unloading steel tubing, quality inspections, packing parts) were analysed by an AI ergonomic assessment tool trialled through a pilot grant from the National Safety Council's MSD Solutions Lab; the tool assigned risk scores by body area and pinpointed the motions most likely to cause injuries (the NSC case study does not name the vendor). At Latham Pools, TuMeke builds a 3D model of the body from the video; trained champions at each site used it to assess tasks and incidents, for example workers bending over conveyor belts to fold products, and to check whether a task could really be causing a reported injury. At Johns Manville, Humantech's motion capture tool analyses force, posture, duration and frequency from a video of any physical task, and a matrix helps the safety specialist choose and prioritize the corrective action.

## Step by step

1. Pull two or three years of injury, first aid and near-miss records and list the sprains and strains by department, task and body part.
2. Pick the ten tasks with the most cases or complaints and record short videos of each with the workers' agreement, explaining why and how the videos will be used.
3. Score the tasks with the video tool and compare the results with what supervisors and workers say is hard; let workers suggest fixes.
4. Fix the highest scores first with engineering changes (height, reach, tool, lift assist, conveyor position) before relying on training alone.
5. Record the same task again after each change to confirm the risk score dropped, and keep the before and after videos for training.
6. Track recordable sprains and strains, days lost and claim costs per site every quarter, and extend to new tasks and new hires' jobs.

## Tools

- AI video ergonomic assessment tool (vendor not named in the case study), funded by a National Safety Council MSD Solutions Lab pilot grant: used at Superior Tube Products
- TuMeke AI ergonomic risk assessment from video: used at Latham Pools
- Humantech motion capture ergonomic assessment from video: used at Johns Manville
- Other tools to evaluate: other video-based ergonomic assessment apps, wearable posture sensors with real-time alerts, the ergonomics module of your EHS software
- To start without new software: your injury log and job list and any LLM (ChatGPT, Claude, Copilot, Gemini) with the prompt below

## Result

Superior Tube Products (National Safety Council case study): 98% reduction in workers' compensation costs and total recordable injury rate down from 13.60% in 2023 to 5.51% in 2024, credited to early identification of risks and the interventions that followed (for example shorter, ergonomic spray wands, which brought an immediate drop in shoulder and wrist strain complaints); assessments took much less time than manual RNLE or REBA studies. Latham Pools (TuMeke case study, the vendor): sprains and strains at its Brantford and Scotia locations fell from 23 cases in 2022 to 2 in 2024 (91%), after workers were taught to let the conveyor bring products to them instead of bending; video assessment also showed that one disputed task was low risk, avoiding an unwarranted claim. Johns Manville (company blog, 2020): program rolled out at 26 plants, analysis time cut from about 45 to 10 minutes per task, 189 ergonomic improvements implemented, and the facilities using the technology reduced their risk of injury by 30%.

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

## Human review: cautions

- The Latham figures come from the vendor (TuMeke) and Johns Manville's from its own blog; the Superior Tube figures come from a National Safety Council case study of a small plant (over 50 people), where a few injuries move the rate a lot.
- The AI only measures risk: the injury reductions came from the changes and training that followed, and other safety work in the same period also counts.
- Filming workers needs their agreement and clear rules on how videos are used and stored; involve them early, as Superior Tube did, and check local privacy law and any union agreement.
- Using video to deny injury claims, as Latham describes, is sensitive: keep the focus on prevention and let a qualified person make claim decisions.
- Video tools estimate posture well but often need you to enter the weights and forces; check the scores against a manual assessment on a few tasks.

## Try it yourself (prompt)

```text
You are an ergonomics and safety specialist in a manufacturing plant. I want to decide which tasks to assess first and what to fix.

I will paste: (1) our injury and first aid log for the last two or three years [date, department, task, body part, type of injury, days lost, claim cost if known]; (2) a list of our main manual tasks [task, department, weights handled, frequency per shift, heights and reaches if known, how many workers do it].

Do this:
1. Group the sprains, strains and other musculoskeletal injuries by task, department and body part, with counts, days lost and cost.
2. Rank the ten tasks to assess first, with the likely risk factors (lifting, bending, reaching, repetition, static posture, force) for each.
3. For each of the top five, suggest engineering fixes first (height, reach, tools, lift assist, layout) and then work method or training changes, with a rough effort level.
4. Draft a short note to workers explaining why we want to film some tasks, how the videos will be used and kept, and how they can suggest fixes.
5. Propose the measures to track per quarter to show whether the program works.

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 use cases

- [AI camera analytics on existing CCTV spot forklift near-misses, missing PPE and speeding so safety teams can coach (Carlex, Epicurean Dairy)](https://aiagentsformanufacturing.com/use-cases/ai-camera-safety-forklift-pedestrian-near-misses-ppe-carlex-epicurean-dairy/)
- [AI video analytics times every cycle on manual assembly lines, so engineers find and fix the real bottleneck (HELLA, a Tier 1 automotive supplier)](https://aiagentsformanufacturing.com/use-cases/ai-video-analytics-manual-assembly-cycle-times-hella/)
- [Connected-worker platform with AI to speed up operator onboarding (Bio-Chem Fluidics, STRONGARM)](https://aiagentsformanufacturing.com/use-cases/connected-worker-ai-onboarding-augmentir/)
- [AI warns paper machine operators before a sheet break and points to the likely cause, so mills lose fewer hours to breaks and rethreading (Sappi, Georgia-Pacific, a Voith paper line)](https://aiagentsformanufacturing.com/use-cases/ai-paper-machine-sheet-web-break-prediction-root-cause-sappi-georgia-pacific-voith/)

## Sources

- https://www.nsc.org/getmedia/115d7d3d-d957-4f76-aa33-339292b12f5c/superior-tube-products-casestudy.pdf
- https://www.tumeke.io/updates/how-latham-pools-reduced-injuries-by-91-and-saved-significant-workers-comp-costs-with-tumeke
- https://www.jm.com/en/blog/2020/september/using-artificial-intelligence-to-evolve-our-safety-strategy/

---
Please cite the canonical URL: https://aiagentsformanufacturing.com/use-cases/ai-video-ergonomic-risk-assessment-sprains-strains-superior-tube-latham-johns-manville/
