AI Developments in Translation & Language Services, curated daily by Anova Translation as part of the AICONTEXT Project.
Industry Intelligence Report
Localization
Interpretation
CAT Tools
MT
Subtitling
Dubbing
9/10
#1 — SlatorCon London 2026 Key Takeaways: 74% of LSIs Now AI-Operational
Executive Summary
Slator published detailed takeaways from SlatorCon London 2026, where 250+ industry leaders gathered on May 22. The headline statistic: 74% of Language Solutions Integrators report being operational, active, or systemic with AI, up from just 6% three years ago. RWS and Cohere presented LanguageWeaver Pro’s architecture — a 100B-parameter model where quality evaluation, infrastructure scale, compliance, and real-time guardrails work together.
Why It Matters
The 6%-to-74% AI adoption jump in three years is the sharpest quantified shift the industry has recorded. For LSPs still in pilot mode, the window for competitive positioning is closing fast. The conference also surfaced a new strategic direction: Deliveroo’s localization team repositioning itself using business-wide metrics rather than translation-specific KPIs, and Acolad’s CTO urging the industry to adopt “banana ketchup thinking” — valuing outcomes over established recipes.
Key Quotes & Details
- RWS Ollie Stott: enterprise translation is about architecture, not just a model — quality evaluation, infrastructure scale, private compliance, and real-time guardrails must work together
- Deliveroo’s Maiju Nurminen: localization teams must “start using the same metrics as the rest of the business”
- Acolad’s Stéphane Cinguino: “banana ketchup thinking” — the outcomes matter, but the recipe can change
- VC panel (Soluna Capital, Arāya Ventures, Twin Path Ventures): building defensible products in a market with giant competitors
- Data-for-AI panel (Appen, Toloka, TELUS Digital): public datasets exhausted, human expertise becoming the key constraint
- Phrase CEO Georg Ell & Tripadvisor: build-vs-buy — don’t reinvent what platforms already offer
- SlatorCon San Francisco announced for September 2026
Key Patterns
Even on quiet news days, the data from the week reveals a clear shift: AI adoption among Language Solutions Integrators has gone from marginal to mainstream in under three years. The 6%-to-74% operational figure from SlatorCon London is the most decisive metric the industry has produced this year. While today’s edition carries a single item, it encapsulates the week’s defining narrative — the translation and localization industry’s AI transformation is no longer a question of if, but of how fast individual companies can adapt.
Watchlist
Tools Gaining Momentum
- LanguageWeaver Pro architecture — RWS and Cohere’s 100B-parameter model with integrated quality evaluation, infrastructure scaling, compliance, and real-time guardrails. Represents a shift from model-as-product to architecture-as-product.
Names to Follow
- Maiju Nurminen (Deliveroo) — advocating for localization teams to adopt company-wide business metrics rather than translation-specific KPIs
- Stéphane Cinguino (Acolad CTO) — promoting “banana ketchup thinking”: focus on outcomes, not recipes
Emerging Themes
- Holistic localization metrics — measuring localization success through business impact, not word counts or turnaround time
- Banana ketchup thinking — the industry’s willingness to adopt unconventional methods when they deliver results
- Data-for-AI as bottleneck — public datasets exhausted; human expertise and curated multilingual data are becoming the key constraint for AI model training
