A major new report from Andreessen Horowitz (a16z) has put some hard numbers on enterprise AI adoption for the first time, and the results are striking.

According to their analysis, 29% of the Fortune 500 and approximately 19% of the Global 2000 are now live, paying customers of a leading AI startup. Not trialling. Not piloting. Live and deployed.

That’s remarkable growth for technology that barely existed in its current form three years ago.

But buried inside this data is an assumption that most organisations haven’t questioned: that AI-powered customer support works when customers speak your language.

Where enterprise AI is winning

The a16z report identifies three dominant enterprise AI use cases: coding, search, and customer support. Of these, customer support is singled out as one of the clearest use cases for demonstrating ROI, because the metrics are simple and measurable. Tickets resolved. Satisfaction scores. Resolution rates. Cost per interaction.

AI agents, the report notes, outperform human agents on every one of these metrics, at lower cost. That’s a compelling case for adoption.

The industries leading the charge? Technology (no surprise), but also legal and healthcare, two sectors historically slow to adopt new software, now moving fast because AI can handle the kind of dense, unstructured text work that traditional software never could.

The gap nobody talks about

Here’s what the a16z report doesn’t address: what happens when your customer doesn’t speak English?

For a Fortune 500 company operating across Asia, the Middle East, or Latin America, deploying an AI support agent that communicates in English solves only part of the problem. The same ROI calculation, more tickets resolved, higher satisfaction, lower cost, breaks down the moment a customer reaches out in Mandarin, Arabic, or Spanish and gets a response they can’t understand.

This is where Babeltext comes in.

Babeltext is a multilingual AI messaging platform that connects customers and organisations in real time across 195 languages. It works across every major messaging channel, WhatsApp, WeChat, SMS, Facebook Messenger, Telegram, and web chat, from a single unified dashboard. Staff work in their own language. The customer communicates in their own. The translation happens instantly and invisibly.

Independent testing by the Ethnic Communities’ Council of New South Wales confirmed a 98% translation accuracy rate across nine major languages.

The complement, not the competition

Babeltext isn’t replacing the AI support tools that enterprises are already deploying. It’s making them work globally.

Think of it as the language layer that sits beneath your existing AI infrastructure. An enterprise can deploy AI agents for support, getting all the efficiency gains the a16z report identifies, while Babeltext ensures those agents can serve customers in São Paulo, Singapore, or Sydney regardless of what language they’re using.

The a16z report notes that the industries succeeding with AI share common traits: text-based work, measurable outcomes, and natural human-in-the-loop involvement. Babeltext is built on the same principles. It blends AI translation with human agents for cases that need empathy or escalation. Every conversation is logged, governed, and auditable, the kind of controls that enterprise buyers require.

What this means for legal and healthcare

Two of the surprise early movers in the a16z data, legal and healthcare, are also industries where language access isn’t just a commercial issue, it’s an equity issue.

A patient who can’t communicate with their healthcare provider in their own language gets worse care. A person seeking legal help who can’t find a firm that speaks their language may not get help at all. Babeltext was built with this in mind; it originated in social impact work, including multilingual mental health support for young people who don’t speak English as a first language.

As AI continues to accelerate in both sectors, multilingual communication becomes foundational, not optional.

The bottom line

The a16z numbers confirm what many in the industry suspected: enterprise AI adoption is real, it’s growing fast, and it’s concentrated in exactly the use cases where AI delivers clear, measurable ROI.

Babeltext sits at the intersection of all three winning use cases, support, search, and omnichannel communication, and extends them to the 195 languages that a monolingual deployment can’t reach.

As enterprises move from pilot to production, the question isn’t just “does our AI work?” It’s “Does it work for all of our customers?”

That’s the question Babeltext answers.

Learn more at babeltext.com