

From Miami
Early in the AI supercycle, we presented our Investment Committee with an opportunity to invest in an AI customer-support chatbot. One of the members, visibly frustrated, warned that it would soon be commoditized — and he was absolutely right.
Indeed, the market is now flooded with general-purpose AI agents that are largely indistinguishable from one another.
It’s no longer just startups like Sierra and Decagon offering AI agent solutions — every major enterprise software company, from Salesforce and Oracle to ServiceNow and Snowflake, now has its own version for customer support, sales and marketing, HR, finance, and more.
Complicating the landscape further, OpenAI and Anthropic are moving into the same territory, embedding configurable agent features directly into their LLMs. And because most startups and enterprise software vendors build their agents on top of those models, the end products have started to look remarkably alike.
Software incumbents claim that customers prefer AI agents that integrate smoothly with the data stored in their core platforms, whether a CRM system or a data warehouse. That “convenience” edge helps incumbents persuade companies to try their AI agents — familiarity sells. Many are even letting customers test these features free of charge.
Looking ahead, as the number of agents inside enterprises explodes, companies will have to decide how to manage them all at scale. The next competition is over the orchestration platforms that govern and coordinate every agent.
But the real shift comes with Agentic AI — systems that don’t just execute tasks but pursue goals. Unlike today’s reactive agents, which depend on human prompts and predefined workflows, agentic systems reason, plan, and act with a degree of autonomy. They connect to tools, APIs, and even other agents to get things done. If AI agents are the apps of this new era, Agentic AI will be its operating system.
Conclusion
Even in the age of AI, competition is the essence — and the beauty — of capitalism.