Already in Full-Scale Commercialization Stage, Speed of Adoption Rather Than Implementation Determines Competitiveness

ElevenLabs Surpasses $600 Million ARR in Four Months This Year, Corporate Valuation Reaches $11 Billion

An analysis has emerged that generative artificial intelligence is evolving beyond the "conversational assistant" stage into a colleague and agent performing actual business tasks for enterprises. Voice AI in particular has already entered full-scale commercialization in enterprise customer response and global content localization, with projections suggesting that for Korean enterprises, adoption speed rather than implementation itself will determine competitiveness.

ElevenLabs held the "ElevenLabs Executive Breakfast Session" on the 3rd at Novotel Ambassador in Gangnam, Seoul.

At the event, Nathan Benaich, Founder and General Partner of Airstreet Capital, delivered a presentation on "State of AI," and Eugene Liang, ElevenLabs APAC Head, shared "ElevenLabs Global and APAC Updates." The moderator and panel discussion were led by Sang-won Hong, ElevenLabs Japan and Korea GTM Lead.

Most notable was the analysis of shifting focus in the AI industry.

Nathan Benaich evaluated that competition in top-tier frontier models has become far fiercer than before, with the US-centric structure being reorganized into a dual-power structure with China.

He particularly diagnosed that the influence of Chinese models in the open-source ecosystem is expanding rapidly.

He explained that the AI industry has moved beyond model performance competition to competition over how precisely to train real-world task environments.

He identified "environment-based learning" as the core driver of AI advancement. Initially, AI learned from games and answer data, but now learns human work methods in spaces that directly simulate software development and research and development environments.

Accordingly, he emphasized that AI is evolving from a simple question-and-answer tool into a collaborator that formulates hypotheses, designs experiments, and validates results.

The speed of real-world industrial application is also accelerating. According to the presentation, the annual revenue scale of OpenAI and Anthropic has grown to approximately $105 billion, and cost efficiency relative to AI performance is also rapidly improving.

In particular, a phenomenon of rapidly widening gaps between top AI-utilizing enterprises and average companies is emerging.

Analysis of US company data showed that top 1% enterprises invested $7,400 per employee per month in AI, while average enterprises spent only about $12. Additionally, enterprise groups actively utilizing AI coding tools showed sales growth rates more than three times higher than low-utilization enterprises.

Changes in the employment market were also a major topic. In fields with significant AI automation impact such as customer support and software development, a decline in new hiring is appearing, but paradoxically, enterprises actively utilizing AI are expanding overall hiring scale, analysis suggested.

Nathan Benaich stated that transferring experienced skilled workers' tacit knowledge to AI will become an important corporate task.

AI security has also emerged as an unavoidable issue.

He warned that the latest AI models have reached a level capable of executing complex cyber attack scenarios, and generative AI-related incidents are expanding beyond simple deepfakes into actual hacking and security threats.

He also pointed out that safety investment scale remains insufficient compared to funds invested in AI model development.

Eugene Liang, who presented next, emphasized that voice AI has already entered real-world service stage.

He stated, "This year is when voice AI becomes reality," and "we have reached a level where it is difficult to distinguish between human and AI voices." ElevenLabs surpassed $600 million in Annual Recurring Revenue (ARR) in just four months this year and recorded a corporate valuation of $11 billion. Over 2 million voice agents are currently deployed on the platform, with over 60% of Fortune 500 companies utilizing them.

He particularly emphasized that the APAC market is showing the fastest growth.

Employment Hero in Australia accessed over 350,000 customers using AI agents and booked over 1,000 demo calls, while Funding Societies, a Southeast Asian financial company, is conducting outbound calls for debt collection and insurance-related matters through AI voice.

Regarding the Korean market, he analyzed that two flows are converging: dramatic demographic structural change and growth in the customer service industry. With declining working-age population yet increasing customer service demand, there is substantial room for voice AI agent utilization, he explained.

Currently, domestic companies such as NH Nonghyup, Shinhan Financial Group, and LG U+ are pursuing automation based on AI agents. Caring, an elderly care company, secured high response rates with an ElevenLabs-based AI well-being call service, he introduced.

During the question-and-answer session at the end of the event, a question was raised: "If you were a Korean corporate executive now, what would you do by year-end?"

Nathan Benaich recommended identifying the best AI use cases already in use by employees and expanding them company-wide.

Eugene Liang advised starting with projects that establish clear use cases and performance metrics and can validate results within 90 days.

Sang-won Hong, ElevenLabs Japan and Korea GTM Lead, stated, "AI is no longer a future technology but already a means of production deployed in corporate operations," and "the difference in competitiveness depends not on the technology itself but on how quickly it is integrated into operations and organization."

He also mentioned, "Voice AI is also entering an inflection point of evolving into a new enterprise operating platform beyond customer service centers and content localization."