Localization Services

No Forrester Wave has ever been published under the name Localization Services, and none has been announced. The evaluation is planned, and the market it would score is already visible through the scorecard Forrester did publish: the inaugural Forrester Wave for Translation Management Systems, Q3 2025.

What the planned scorecard would have to score

Localization services is the human half of a market whose technology half just got its first evaluation. The services layer is what the language service providers actually sell: translation and localization delivery, linguistic quality management, program management, and the increasingly thin margin between machine output and human review.

The planned Wave, whenever it publishes, would score firms whose business model is being rewritten by the same AI that the technology Wave already grades. That tension is the article's subject, because it is the market's subject.

The technology scorecard that got here first

Principal analyst Kathleen Pierce announced the inaugural Translation Management Systems Wave on September 10, 2025, scoring twelve vendors on nineteen current offering criteria and seven strategy criteria.

The field matters for the services buyer because it contains the platforms owned by the services giants: Lionbridge, RWS, and TransPerfect are all scored, alongside Bureau Works, Centific, LILT, Lokalise, memoQ, Phrase, Smartling, Unbabel, and XTM. TransPerfect's acquisition of Unbabel was announced just before publication, and the two were still scored separately.

Three Leaders surfaced publicly. Phrase, with the highest possible scores in twenty one of twenty six criteria, on configurable workflows, continuous localization, and AI agents that automated ninety nine percent of workflows for some reference customers. XTM, on orchestration and AI agent roadmaps, workflow controls, and fit for regulated industries. LILT, on custom AI models paired with human-verified precision.

Pierce's own framing is the one to carry into any services evaluation: "AI reshapes localization workflows, not just translations." And the sharper version: "The real uplift comes from workflow automation and agents, and TMS providers are not at all standardized in this respect."

The services layer under the technology layer

If the TMS is, in Pierce's words, a "vendor-neutral orchestration layer," the services providers are the labor and judgment that run on top of it. The technology Wave already reveals the services market's squeeze from both sides.

The upside: automation raised what one account team can deliver, and the Leaders' numbers, ninety nine percent workflow automation, are services economics as much as software features. The downside: integration is the backbone of modern TMS platforms, and Pierce notes that many reference customers wanted to switch providers but were held back by integration investments. Lock-in in the technology layer flows directly into the services layer, because the services contract usually rides on the platform contract.

There is also a second published signal for the services market. Forrester's Wave on Customer Experience Strategy Consulting Services, Q4 2024, found that localization and language competencies are a major differentiator for consulting firms, with clients citing the difference between culturally competent translations and straight translations. That distinction, cultural competence versus literal accuracy, is precisely the services value proposition no software scorecard can measure, and the closest published proxy for what a Localization Services Wave would score.

What a services Wave would have to decide

A future Localization Services Wave faces a definition problem the technology Wave did not. What is the unit being scored? The linguist network? The AI orchestration layer? The account management that survives a merger? The three announced TMS Leaders answer differently: Phrase sells the automation, XTM the governance, LILT the custom-model-plus-human mix.

The services firms in the same field answer differently still. TransPerfect just bought Unbabel before the Wave published, the classic services-industry move, consolidation to buy the technology margin back. A services evaluation published today would be scoring an industry mid-consolidation, which is why the planned status may be the honest status.

There is also the pricing problem a services scorecard cannot dodge. Machine translation compressed per-word rates for years, and now AI agents are compressing the workflow costs around the translation itself. A services evaluation that scores firms on the old economics, cost per word, will flatter the incumbents. One that scores on the new economics, cost per outcome across an automated pipeline, will reward the firms that rebuilt. Which economics Forrester chooses, whenever the Wave publishes, will decide whose Leader status means anything.

The honest state of this category

The planned evaluation does not exist yet, and this article does not pretend otherwise. What exists is the published technology scorecard, the CX consulting cross-reference, and the vendor announcements, and that is what this article works from.

A buyer shopping localization services today should read the situation as it is: the technology half of the market is now scored, the services half is not, and every services vendor is racing to look like a technology vendor before the scorecard catches up. The gap between the two halves is where the value currently sits, and where the next scorecard will land.

Three questions while the Wave is pending

Which half of the deal are you actually buying: the platform or the people? The TMS Wave scores the platform. Nothing yet scores the people. Separate the two in the contract, because the vendor will not separate them in the sales cycle.

What is your automation number, and what does it mean? Ninety nine percent workflow automation is a platform claim. Ask your services partner what percentage of your content still requires human review, and what that review costs per word, in your language pairs.

Who owns the integration when the contract ends? The lock-in finding cuts both ways. Ask what happens to your workflow automations, glossaries, and translation memory when the services contract ends, and get the export paths in writing. A services firm that cannot answer this is already planning to keep you through inconvenience.

Forrester does not endorse any vendor named here, and tier placement should not be read as a recommendation to buy.

A close neighbor in Forrester's own coverage is Marketplace Development Platforms, where this Wave scores one category built from four genuinely different businesses, selling physical goods to consumers versus digital services to businesses, and its two Leaders each dominate one end of that ladder and barely compete at the other.