Kevin O’Donnell, founder of Global10x and former VP of International Growth at Dropbox, joined Jason Hemingway on the In Other Words podcast to talk about where AI is creating genuine new value in international markets, why hyper-localization changes the game, and where human judgment still matters.
The AI conversation in international growth tends to center on efficiency. Faster translation. More volume. Lower cost per word.
Kevin O’Donnell thinks that conversation misses the point entirely.
The genuinely interesting opportunity, the one leadership teams should be paying attention to, is in use cases that never existed before.
New audiences, not faster output
Kevin’s argument is that AI’s real value in international growth is making it economically viable to reach audiences that were previously too expensive or too niche to serve. It is the ability to create original, culturally relevant content for a particular demographic in a particular market. Content that didn’t exist before because the cost of producing it through traditional methods could never be justified.
“In many ways, this is not replacing anything. It is net new. In the past, there was zero content for these audiences.”
What hyper-localization means
Kevin calls this hyper-localization. Hyper-localization is the practice of creating original content shaped by an audience’s cultural context, demographics, and local market conditions, rather than adapting existing content from another language.
The example he returns to is content for students in Barcelona who speak Catalan. Not adapting a Spanish-language page for a Catalan audience, but generating content built around who they are, where they live, and what matters to them.
“It’s not just localizing content in Catalan, but actually going very niche into different demographics in different regions. That’s brand new and that’s compelling.”
Until recently, producing this kind of highly targeted content at scale was economically unrealistic. The audience size was too small and the production costs too high.
AI removes much of that constraint, making it possible to serve smaller and more targeted audiences that companies would previously have ignored.
Why hyper-localization is no longer optional
Kevin sees hyper-localization moving from competitive advantage to baseline expectation, particularly for marketing teams. The ability to reach specific audiences with meaningful relevance at scale will become a requirement for companies serious about international growth.
“I think it is going to become a requirement for more teams to do this, particularly marketing teams. They have the means to have this global reach, not just in terms of scale, but now in depth and breadth.”
The implications extend far beyond marketing. Product teams can localize onboarding experiences market by market. Support organizations can finally justify help content for lower-volume languages. Sales teams can generate market-relevant collateral at scale. Audiences that were previously too small or too expensive to serve are now commercially viable.
The companies that move first on hyper-localization will build relationships with audiences their competitors haven’t even identified yet. And once those relationships are established, they become genuinely difficult to displace.
Where human judgment still matters
Kevin pairs the opportunity with a clear caution. High-stakes content, anything where the consequence of getting it wrong is reputational, legal, or financial, still demands human oversight.
The obvious categories are legal, compliance, medical, and financial content. But Kevin adds homepages, pricing pages, and any brand-defining moment to the list. These are the places where trust is earned or lost, and where the cost of error goes well beyond a single piece of content.
“High-stakes content that needs perfect quality is usually where the consequence of error is really high. So there might be reputational damage, legal exposure, and lost revenue.”
The companies getting the most value from AI are not eliminating human oversight entirely. They are building content strategies that distinguish between high-, medium-, and low-risk content and applying governance accordingly.
That allows teams to accelerate aggressively where speed and scale matter most while protecting the moments where accuracy, trust, and brand integrity carry higher stakes.
Scaling hyper-localization
This is where infrastructure becomes critical.
Hyper-localization only works at scale when organizations can manage multilingual content, AI workflows, human review, and governance in a connected system rather than fragmented tools and regional workarounds.
Intelligent platforms such as Phrase are increasingly designed around that challenge, combining AI-powered content creation with quality management, workflow orchestration, and human oversight where it matters most.
The companies that benefit most from AI in international markets will not simply produce more content faster. They will understand where automation creates advantage, where human expertise still carries outsized value, and how to build an operating model that supports both at scale.
Watch the full conversation with Kevin O’Donnell on In Other Words.
