Are established customer service platforms running out of time to build truly AI-native customer experiences from within?
In this CX Today interview, Nicole Willing speaks with John Kim, Co-Founder and CEO of Delight.ai, about Salesforce’s acquisition of Fin and what it reveals about the next phase of competition in customer service AI.
The deal is less about one acquisition and more about a broader market shift, Kim says. Legacy customer service platforms are under pressure to move faster as AI-native companies reshape what enterprise buyers expect from automation, support workflows and customer engagement. Rather than relying only on internal product development, some of the largest vendors are now looking to acquire AI-native capabilities that can help them compete at the speed of the new market.
The window to lead in AI-powered customer service is closing quickly, particularly as AI agents move from experimental pilots into live production environments, Kim notes. Traditional customer service platforms may have scale, brand recognition and enterprise relationships, but AI-native providers often have a different advantage: they are built around AI from the start.
“What we’re seeing here is this first generation, which used to be the traditional enterprises, whether it be Salesforce or Zendesk, trying to compete on their own, ultimately realizing that they’re not fully able to compete at the same rate as the AI-native companies. So now they’re taking this strategic approach to acquire them so that they can actually compete at the same velocity.”
That difference in velocity matters. Kim explains that AI-native companies are usually structured to test frontier models, optimize accuracy, reduce latency and lower the cost of service delivery more aggressively than legacy platforms. They are also less constrained by existing seat-based business models, which can make it harder for established vendors to fully embrace automation that reduces the need for human agent licenses.
The discussion also explores what this means for CX leaders and enterprise buyers. Kim cautions against both overestimating and underestimating AI. Some organizations expect AI agents to solve every service challenge immediately, while others judge today’s tools by the limitations of older chatbot systems. Both views can slow progress.
Instead, Kim encourages CX teams to start with focused, narrow use cases in production, where they can learn from real customers, real traffic and real operational outcomes. That approach helps leaders build a more practical understanding of where AI can improve response times, reduce customer effort and support service teams without adding unnecessary risk.
Salesforce’s acquisition of Fin points to a customer service market where AI-native capability is becoming a strategic necessity. For CX leaders, the challenge is how quickly their organizations can prepare to use agents safely, effectively and at scale.
