Published on August 29, 2025
Organizations have invested billions in generative AI. Yet, according to a recent MIT report (State of AI in Business 2025), 95% of generative AI (GenAI) projects have produced no tangible return on investment. Adoption is massive — nearly every company has tested ChatGPT, Copilot, or similar tools — but the actual value created remains limited.
The translation industry is facing a similar reality. Scriptis has been testing various AI technologies for over a year, but results have been underwhelming. Most companies in our field have reached the same conclusion (though it’s not the end of the road). In late 2024, many believed that feeding large language models (LLMs) with increasingly extensive datasets would improve pre-translations. In 2025, however, a growing consensus among experts is that AI has reached a performance ceiling.
The MIT report highlights two main levers for creating real value with artificial intelligence:
✅ Prioritize partnerships: initiatives carried out with external providers are twice as successful as in-house developments.
✅ Focus on internal processes: the biggest gains come from automating finance, procurement, and customer service rather than sales & marketing projects.
What does this mean for organizations that frequently use external translation services?
☑️ AI is not a magic wand: generic tools like ChatGPT cannot guarantee terminological consistency, regulatory compliance, or brand integrity.
☑️ Integration drives value: real gains occur when AI is embedded in established workflows. Achieving AI-driven efficiency in translation without sacrificing quality requires expertise in translation memories, quality control, cultural adaptation, and style guide compliance.
☑️ Partnership matters more than tools: working with an experienced translation firm enables companies to combine AI with human expertise — lowering costs, accelerating time-to-market, and ensuring quality.