At FT Strategies, we have helped over 100 publishers identify AI opportunities based on value and feasibility, and one use case we are seeing across many organisations—whether it’s a news publisher, broadcaster, or book publisher—is AI-powered content translation. This is a robust strategic bet as it has the potential to efficiently unlock audiences in new regions and expand overall content reach. Use cases like this, which deliver significant and measurable value, are becoming increasingly important to identify as organisations move beyond early experimentation.
To understand this use case in more detail, we spoke to Akshat Prakash, Chief Technology Officer at Camb.ai — a speech and translation AI tool with proprietary models capable of dubbing content into over 140 languages. We initially approached the conversation to understand how Camb.ai’s models differ from other industry-standard tools like ChatGPT or DeepL. But as we spoke with Akshat, it became clear that the real story wasn’t just about the technology — it was about how enterprises are navigating the different economic choices around this use case, and overall, how they are making AI investment decisions.
The broader question: how to ensure AI use cases bring ROI?
AI investment decisions are often guided by a strategic direction set at the top, often a CEO or Board recognising AI’s potential and seeking measurable impact. However, quantifying the returns of AI use cases is challenging because benefits emerge as workflow efficiencies and product innovations, which may not directly translate into financial outcomes. While some AI applications, such as dynamic pricing or programmatic advertising, yield clearer ROI, others generate impact in ways that are harder to isolate and measure. And the challenge is not just measurement but attribution - determining AI’s unique contribution amidst broader business changes. To make AI investment decisions more tangible, organisations need a structured way to assess and prioritise use cases that balances quick wins with more ambitious projects, supported by on-the-ground measurement to validate impacts.
Akshat’s approach to this challenge is to develop a spectrum of use cases. He will sit down with organisations, draw a line, and start laying out use cases which can be iteratively and intuitively compared with one another in terms of complexity, risk and cost. On one end of the spectrum are “use cases which should have been started yesterday” - low-risk, straightforward implementations where AI reduces costs and streamlines workflows, such as transcription tools for journalists, tightly-scripted customer support automation or simple 1:1 translation tools. On the other end are more transformative opportunities, such as AI-powered multilingual video dubbing or real-time broadcast translation, which require greater investment but unlock new revenue streams and audiences. Akshat shared a recent example of Camb.ai’s work where this consideration was particularly relevant: a 2+ hour Major League Soccer livestream broadcasted in four languages simultaneously, including Bengali, which represents a huge amount of audiences in a market that is nonetheless relatively untapped by translated content. Through this exercise, organisations can uncover low-hanging fruit AI use cases that they should already be implementing. This structured approach not only highlights quick wins but also helps prioritise more ambitious innovations over time.
Bringing it to life: AI-powered translation and Camb.ai
When organisations think about “reaching audiences in new languages”, they are not necessarily aware of this variety of approaches. But Akshat explained with his personal experience: having grown up internationally and speaking English fluently, he nonetheless experienced a disconnect when coming to the US and finding it difficult to understand a coffee shop cashier. That moment stuck with him — it showed him that language barriers aren’t just about different languages but also about accents, speed and accessibility. Having worked on early AI-driven language models for Apple’s Siri, and later developing a tool to help students fight employment, Akshat saw firsthand how AI could be used to bridge gaps in communication and access. These experiences ultimately shaped Camb.ai’s mission.
“The internet was created for English speakers, it's 2025 we should redesign it for the world”
The platform allows organisations to translate and dub all types of content—whether the content itself or the accompanying materials—across multiple languages, from text-to-text and text-to-speech.
Akshat highlights two key factors for AI-powered translation. Firstly, authenticity is crucial. Organisations might jump into voice cloning because of its novelty, but that doesn't necessarily mean it’s what your audiences want. He points out that what audiences truly want is for the voice to sound natural and authentic, blending in cultural nuances, idioms and accents. This is where localisation becomes crucial — not just translating words, but ensuring tone, pacing and phrasing feel native to each audience. At FT Strategies, we have seen the importance of this when working with news publishers, including an engagement with El País, where we tested the idea of producing AI-generated podcasts. Authenticity — measured by how natural and true-to-life the AI-generated Spanish translation sounded compared to the host journalist’s actual voice — was key in evaluating the initiative's success and determining next steps. The focus groups that we ran showed that readers enjoyed the podcasts and perceived the voices to be believable and authentic.
Secondly, it is important to look beyond your content and at your wider brand and its perception in the new market. Akshat notes that when organisations aim to engage every fan in their native language, “it's not just about translating content [to different languages]; you need all engagement assets —PR, marketing, and beyond—available in multiple languages.” Without considering the full distribution process, publishers could miss out on reaching and engaging their new audiences to the fullest extent. This holistic approach requires a flexible and scalable infrastructure, which Camb.ai structures across three layers. At the foundational level, proprietary AI models - BOLI (speech-to-text) and MARS (text-to-speech) - provide the core translation. As this layer becomes widely commoditised, the real advantage will come from how it’s packaged into a product rather than simply building the models. An infrastructure layer scales this to serve traffic across multiple continents, ensuring global accessibility. Finally, the product layer (i.e., the packaging of the foundational and infrastructure layers) enables multi-format adaptation, refining AI outputs for different mediums — whether live broadcasts, written content or marketing materials — while ensuring natural, authentic tone and delivery.
Not every publisher will need a technology stack of this level of sophistication. For example, while working with a European book publisher, we found that many off-the-shelf tools could support multi-format content creation, offering a low-cost, low-risk solution. However, publishers can still assess and reflect on the role each layer plays in their business and tailor their approach to AI accordingly.
Conclusion
AI-powered translation is more than just a tool for breaking language barriers—it is a strategic enabler for audience expansion, brand positioning, and long-term growth. However, success depends on more than just deploying AI models; it requires localisation, editorial oversight, and scalable infrastructure to ensure translations resonate authentically with new audiences. Le Monde’s English edition illustrates this well. Rather than relying solely on AI-generated translations, Le Monde combines machine translation with human editorial review, ensuring articles maintain their journalistic integrity and cultural relevance. This balanced approach has driven a 10% monthly audience increase, with visits surpassing 3 million since its launch in April 2022.
As AI models become increasingly commoditised, the competitive edge will come from how organisations integrate and apply AI within their workflows. The true differentiator lies in selecting the right use cases, aligning AI investments with business objectives, and ensuring that AI delivers measurable impact, rather than just technological novelty.
At FT Strategies, we understand the power of AI, technology & data for future-proofing your business. Our specialist team of consultants have years of experience implementing data models and refining tech stacks to help organisations grow. To find out more about our services, please get in touch with us today.
About the authors
Sam Gould, Manager, AI Lead
Sam is the AI Lead at FT Strategies with over 7 years of experience helping clients solve strategic business challenges using data. He has helped organisations in both the public and private sectors to define strategic roadmaps and processes for using AI. He has also designed and built innovative data solutions, working with senior stakeholders as part of critical delivery-focused teams.
Inés Luque Calatayud, Associate Consultant
Inés Luque is an Associate Consultant who previously worked in the Business Development department of the Financial Times. She has experience working on subscription funnel analysis and strategic direction support. Inés holds an MSci in Management Science from University College London.
