Challenge: Siloed teams, fragmented processes, and zero consistency

When Rosario Messina, Senior Product Operations Manager, joined Awin the company's localization processes were inefficient.

Six teams were managing 1.5 million source words annually across eight languages (German, French, Polish, Dutch, Italian, Swedish, Spanish, and Portuguese), using spreadsheets and various other tools while working in silos. 

In total, Awin was translating about 12-13 million words a year.

1. Marketing team generating ~50% of content using various tools
2. Platform services managing advertiser-related content through various tools
3. Engineering localizing the platform code
4. UX and product teams working independently on user-facing content
    

We had situations where some teams were leveraging a mixture of work management tools like Asana and DeepL translation, while others were just going to ChatGPT, translating content, and rolling out requests through Word documents or Excel files.

 

Rosario Messina

Senior Product Operations Manager at Awin

This decentralized approach created significant pain points across the organization.

Each team used different translation approaches, styles, and terminology, leading to inconsistent content, duplicate translation efforts and the inability to leverage shared translation memory across teams.

Awin needed a part-time translation manager per team. Depending on the team turnaround times were slow — 4 weeks from request to delivery — which didn’t scale.

All of this needed to be resolved without increasing their existing budget.

Solution: One platform where every team collaborates

Rosario identified the pain points early on and worked backwards to determine each team’s requirements.

Why Lokalise?

We were looking for a system that would become our single source of truth. We needed task management, comments, AI with a human approach, and something that worked really well with our platform. So a strong API implementation. Lokalise ticked all those boxes.

 

Rosario Messina

Senior Product Operations Manager at Awin

• Strong API capabilities for platform integration
• Task management integration for contextual workflow management
• Comments enabling real-time collaboration on specific translation segments
• AI capabilities ready for human-in-the-loop workflows
• Existing engineering adoption providing a foundation to build upon

As for Acclaro, Awin was looking for a language partner who had past experience using Lokalise and who was already working with companies of their scale. Previous agency relationships had failed due to Awin’s niche industry and specialized terminology.

Acclaro brought three critical advantages:

1. Localization tech expertise: Experience working within Lokalise
2. Enterprise scale: Proven track record with companies like Netflix, Meta, and Amazon
3. Flexibility: Ability to adapt to Awin's unique industry requirements and processes

What we found with Acclaro is a translation agency that showed from the very beginning they can adapt to whatever the requirement is, which is obviously very important, especially in our industry as we are pretty niche.

 

Rosario Messina

Senior Product Operations Manager at Awin

Implementation: AI + human expertise at scale

Awin's implementation centered on creating a unified workflow that leverages both AI efficiency and human expertise.

The team configured Lokalise with their existing translation assets — glossaries, style guides, and historical translation memory — making sure that AI translations were tailored to their specific terminology and brand voice.

They established standardized request processes through Asana forms with automated routing to their dedicated translation manager, eliminating the confusion of multiple contact points.

Once content enters Lokalise, the AI engine translates it using the context and linguistic assets available, delivering contextually accurate content.

Acclaro linguists then receive tasks with complete context through Lokalise's integrated workflow, performing machine translation post-editing (MTPE) to refine accuracy, ensure brand consistency, and adapt cultural nuances.

Once the main translation task is completed, chained tasks automatically notify Awin teams that content is ready for internal review..

“It takes us one or two days to get from source text to post edited by humans. And then we try to give a week to our colleagues to review.”

Once translations are complete, content is automatically delivered, either as document downloads for marketing materials or via the GitHub integration for platform content.

Integrations and tools used by Awin

• GitHub: Seamless platform localization for engineering content
• Asana: Ticketing system for localization workflow management
• Translation memory: Continuous learning from past translations to improve time, accuracy, and costs
• AI orchestration + linguistic assets: Glossaries and style guides are used to inform AI’s decision on which large language model delivers the best translation

Results: 57% speed increase and higher quality translations without budget increases 

The first visible improvement was in delivery speed, with turnaround times dropping dramatically.

At the beginning, for a marketing translation, it took us on average about four weeks from request received to localization delivered. Today, we are down to less than 12 days — that's an increase in speed of 57%.

 

Rosario Messina

Senior Product Operations Manager at Awin

The team's target is even more ambitious: 2-5 day turnarounds through a fully automated process.

Beyond speed improvements, Rosario has even been able to cut costs equivalent to the salary of three full time employees. He has used this budget towards hiring a full time translation manager, Isil Serbes Aliprandi, who has been a game changer for the program, and centralized localization process:

"We used to rely on seven team members who handled translations as a side task. Today, we’ve streamlined the process with one specialist, completely removing the translation workload from those teams."

For optimized languages, internal reviewers spend dramatically less time on post-translation QA. 

We have seen a decrease in review time that ranges between 70 and 80% for regions where quality and consistency of translation are meeting our standards.

 

Rosario Messina

Senior Product Operations Manager at Awin

In short, Awin has:

• Centralized all localization assets in Lokalise
• Achieved standardized processes across all teams and content types
• Improved quality and speed, without increasing budgets

Future: Automation and optimization

Awin's localization transformation is ongoing, with clear next steps:

In the short-term, Rosario and Isil would like to eliminate internal review for content that doesn’t pose a reputational risk. Currently, the internal review takes the longest. He’d also like to expand API integrations with content management systems and optimize style guides and glossaries for other languages.

In the future, Rosario would like to set up a fully automated workflow for standard content types.

"Our objective is to be able to turn translation around between two and five days, which is absolutely possible through automation and leveraging all of what we have within the system now," explains Rosario.

Key takeaway: Foundation first, technology second

When asked for advice for companies starting their localization journey, Rosario emphasizes fundamentals: 

"Get your translation assets right and early — glossary, style guides, quality translation memory. Based on that, AI delivers better translations, linguists have an easier job, and there's less time required on aligning styles."

The Awin transformation demonstrates that with the right platform, expert partners, and systematic approach, you can achieve enterprise-scale efficiency and quality.