Language AI has changed what businesses can localize, who can do it, and where it happens. It has also raised an obvious question: Do enterprises still need localization service providers when translation is being built into everyday applications?

Last week, we published The Forrester Wave™: Localization Services, Q3 2026 to help answer that question. We evaluated 11 leading providers out of the hundreds that exist, focusing on established localization firms that enterprises have relied on for years. I wanted to understand whether these providers are adapting to the age of AI, what they can now help customers accomplish, and how buyers should assess them.

The short answer is yes, enterprises still need LSPs. But they should expect a very different relationship.

Here are three tips for using this Wave and seven questions that can help you start changing that relationship now.

Tip 1: Don’t Stop At The Graphic

Most people read the report, look at the Wave graphic, and decide that they have seen the evaluation. They have not.

Use the “Customize Your Shortlist” functionality on our website or download the Excel model and open the Criteria tab. In my estimation, that page contains about two-thirds of the value of every Wave. This evaluation includes 127 questions across 27 criteria covering what a modern LSP should be able to do, from orchestrating AI-enabled workflows and governing self-service translation to curating linguistic assets, supporting regulated content, operationalizing cultural intelligence, and measuring business impact.

The scoring scales are just as important as the questions. They describe what below-par, on-par, and above-par capabilities looked like among the providers we evaluated. We developed those scales after reviewing the providers’ responses, listening to their strategic briefings, seeing their demonstrations, reading their case studies, and speaking with their customers.

A Wave is closer to the Olympics than an open tryout. Every participant is already a significant competitor that Forrester clients work with and ask about. The scoring is relative to that highly capable field, and the scales reveal what separated good capabilities from the best we saw. That makes the Criteria tab useful well beyond these 11 providers. Use it to prepare an RFP, evaluate a renewal, challenge an incumbent, or decide what capabilities your organization will need next.

Tip 2: Read This As A Services Wave

Technology mattered throughout the evaluation, but this was not a comparison of platforms and features. We looked at how providers combine technology, services, human expertise, data, and operating infrastructure to help customers achieve business goals.

That distinction matters. A services provider does not necessarily need to build or own every piece of technology it uses. It does need to know which tools and models to deploy, how to connect them, where to put controls and human judgment, and how to keep the entire system working as the market changes.

This Wave asks whether providers do that well.

Tip 3: Choose For Fit, Not Position

In the first blog I wrote at Forrester, I compared choosing a localization service provider to dating. Someone can look perfect on paper and still be wrong for you. Another provider may not seem like the obvious choice but may understand your business unusually well. That remains true.

Any provider in this Wave could be an excellent fit for one enterprise and entirely wrong for another. A global government entity or enormous corporation may need the infrastructure, geographic reach, and depth of resources available from a large provider. Another organization may be better served by a provider with particular domain expertise, a more compatible culture, or a delivery model that matches how its teams work.

The graphic tells you how LSPs performed against our criteria. It does not eliminate the need to understand how they map to your own requirements.

Seven Questions To Ask Your LSP Now

The providers showed a high overall level of capability, and what separated them was not simply whether they could perform a given service. With enough time and money, any services firm can usually say yes. The more useful question is whether the provider has turned its experience into repeatable capabilities that anticipate how localization is changing and help its customers move forward with skill and judgment.

Here are seven questions I would ask:

  • Have you turned exceptional projects into repeatable capabilities? An impressive custom solution proves that smart people solved one problem. Look for methods, tools, decision rules, and feedback loops that allow the next customer to benefit from what they learned.
  • What sort of infrastructure have you built for regulated and sensitive work? Look beyond general industry experience. Ask about specialized teams, validated processes, auditability, jurisdiction-specific requirements, and infrastructure for areas such as clinical trials, patents, legal filings, and regulated product content.
  • Can you govern localization where people actually work? Localization is always decentralized across diverse teams and locations. The mandate is not to centralize but to coordinate, and provide the right level of rigor for the risk. Providers should help you uncover “shadow” translation and bring standards, model controls, and auditability into the applications employees already use.
  • Do you understand the new risks created by multilingual AI? Standard privacy and security controls are only the beginning. Ask how the provider evaluates models across languages, manages data and model choices, tracks changing regulations, tests outputs, and preserves human accountability.
  • Can you make your multilingual content useful to AI systems? Search is becoming a multilingual knowledge problem. Ask how the provider supports answer-engine optimization, structured content, multilingual taxonomies, knowledge graphs, retrieval, and measurement of brand visibility across languages and locales.
  • Can you turn cultural expertise into reusable intelligence? In-market experts remain essential, but their knowledge should not disappear into one-off reviews. Ask how the provider captures cultural, demographic, behavioral, and market insights so that they can inform future content, campaigns, AI systems, and business decisions.
  • Can you test the whole experience, not just the words? Multilingual quality failures increasingly appear in interfaces, images, video, conversational systems, and multimodal experiences. Providers should be able to find layout, truncation, usability, cultural, functional, and visual problems before customers do.

LSPs May Be Headed For A Renaissance

For years, people have predicted that automation and AI would steadily make localization service providers less relevant. On the contrary, they make localization expertise essential.

LSPs have spent decades learning how language and communication vary across markets, when automation requires human judgment, how to manage ambiguity and validate specialist content, and how to find and vet expertise around the world. As enterprises deploy AI across languages, demand for that knowledge and capability will rise.

The organizations that get multilingual AI right will not simply turn on translation and walk away. They will define requirements based on organizational objectives; clean and structure upstream data; test multimodal production scenarios; establish risk-based workflows; and keep optimizing as models and business needs change. A strong LSP can help with all of that.

My conclusion from this Wave is not merely that I would still use a localization service provider. I would take a fresh look at the entire relationship. I would ask what the provider knows that my organization does not, where it can supply infrastructure or expertise that we cannot reasonably build, and how it can help us move from buying translated words to building a multilingual operating capability.

Read The Forrester Wave™: Localization Services, Q3 2026. Then use our website to customize your shortlist or download the Excel, study the criteria and scales, and use them to decide what you should expect from your LSP now.

Share