Increasing translation output while reducing effort
As a global manufacturer serving customers in more than 100 countries, Ariel needed to scale translation output without increasing cost or manual effort. By extending its XTM and Adobe Experience Manager setup with machine translation and neural fuzzy adaptation, the team reduced post-editing time and unlocked more value from its language assets.
Case study
How Degreed reduced linguist questions by 80% using visual context.
About Ariel
Ariel Corporation is a family-run American manufacturer founded in 1966. What began in a basement has grown into the world’s largest manufacturer of separable reciprocating gas compressors, while retaining its entrepreneurial ethos.
Today, Ariel employs more than 1,500 people across multiple locations in Ohio. Its compressors are used to extract, process, transport, store, and distribute natural gas in more than 100 countries worldwide.
To support global customers, Ariel translates its website, support portal, video content, technical documentation, and online learning materials from English into Spanish, Chinese, and Russian.
Industry: Oil and gas manufacturing
Founded: 1966
Headquarters: USA
Markets: 100+
The challenge
Scaling output while reducing localisation spend
Ariel already had a strong localisation setup in place. Adobe Experience Manager was connected to XTM through a connector, allowing content to flow automatically between systems. This meant the localisation team spent minimal time managing operations.
While the workflow was efficient, most translation and review work still relied heavily on human effort. As content volumes increased, Ariel wanted to increase translation output, reduce post-editing time, and lower overall localisation spend without sacrificing quality.
Key localisation goals included:
-
Reduced post-editing time
-
Increased translation output
-
Lower localisation costs
To achieve this, Ariel began exploring whether machine translation could take on a larger role in its workflow.
Why XTM
Building on an automated localisation foundation
Rather than replacing existing systems, Ariel wanted to build on what was already working. XTM provided a stable platform that integrated tightly with Adobe Experience Manager and could support more advanced automation.
XTM’s ability to manage translation memory, route content automatically, and integrate with machine translation engines allowed Ariel to introduce new efficiencies without disrupting established workflows. This made it possible to improve performance incrementally while maintaining control over quality.
The solution
Combining machine translation, automation, and human review
In 2020, Ariel integrated SYSTRAN machine translation into its existing XTM and AEM workflow. Content flowed automatically from AEM into XTM, where translation memory was checked first. If no perfect match was found, the content was sent to SYSTRAN for machine translation.
Human translators then post-edited the machine output, followed by review from regional officers. After quality checks, the final translation was automatically returned to AEM for publishing.
This workflow reduced manual handling and improved consistency across content types.
When XTM released version 12.7 with neural fuzzy adaptation (NFA) in 2021, Ariel saw another opportunity to reduce effort. NFA learns from previous translator corrections and turns fuzzy matches into full matches, meaning fewer edits were required, and translators could move faster.
Ronald Egle, Content Systems Administrator at Ariel, explains:
“By using XTM and SYSTRAN, we’ve been able to find extra value in our language assets. We had a phenomenal setup with our content management system feeding automatically into and out of XTM. Adding SYSTRAN MT and then neural fuzzy adaptation, we found gold. Now we’re able to harness the very best of both machine translation and human editing for outstanding cost efficiency.”
Higher efficiency, lower effort, measurable savings
The impact of combining XTM, SYSTRAN MT, and neural fuzzy adaptation was clear and measurable. Machine translation quality improved by 100 percent, while human translation effort dropped by 31 percent. Post-editing became faster, and overall translation output increased.
Before the introduction of NFA, machine translation accounted for around 17 percent of Ariel’s total translations. After implementation, the share of human translation dropped significantly, allowing more work to be completed with less effort.
Translators now receive stronger machine-generated suggestions, which improves speed and provides high-quality reference material. Reviewers benefit from greater consistency across technical terminology and product content.
Improvement in MT quality
Reduction in human translation
Increase in translation output
Foundation for future optimisation
With a more capable machine translation setup in place, Ariel is now looking ahead. The localisation team is working closely with content creators to structure source content using more predictable language and patterns, allowing machine translation to take on a greater share of the workload over time.
This forward-looking approach aligns with XTM’s continued platform development. In 2025, XTM introduced advanced AI capabilities, including agentic AI tools designed to support deeper automation, smarter decision-making, and more adaptive localisation workflows.
By combining disciplined content creation with a platform that continues to evolve, Ariel is well-positioned to take advantage of future automation and AI-driven improvements as they become available.
The result is a localisation foundation that supports efficiency today while remaining ready for what comes next.
See how XTM supports machine translation workflows
Start a free trial to explore XTM in your own localisation setup, or request a demo to discuss MT integration with a specialist.
FAQs
How did Ariel reduce post-editing effort without losing quality?
Ariel combined machine translation with structured human review rather than replacing linguists entirely. By integrating SYSTRAN with XTM and introducing neural fuzzy adaptation, translators received better suggestions and spent less time editing while maintaining accuracy.
What role did neural fuzzy adaptation play in Ariel’s workflow?
Neural fuzzy adaptation learns from previous translation corrections and converts fuzzy matches into full matches. This reduced manual editing and improved consistency, helping Ariel increase output without increasing effort.
How does XTM integrate with Adobe Experience Manager?
XTM connects with Adobe Experience Manager through a connector that enables automatic import and export of content. This allows localisation workflows to run with minimal manual intervention, freeing the team from day-to-day operational management.
Why didn’t Ariel replace its existing localisation systems?
Ariel already had an effective automated workflow in place and a TMS solution in place with XTM. XTM allowed the team to extend that foundation with machine translation and automation rather than starting over, reducing risk and disruption.
Is XTM suitable for technical and manufacturing content?
Yes. The Ariel case study shows how XTM supports technical documentation, product content, and learning materials by combining automation, translation memory, and human review to maintain accuracy at scale.
Speak to our team to find the right solution for your business.
Resources for scaling machine translation
Explore guidance on combining machine translation with human review, getting more value from language assets.
Ready to scale translation output without increasing cost?
See how XTM helps teams automate localisation, improve machine translation quality, and reduce manual effort.