In a new joint post, multiple a16z partners predict that the record-system model dominated by Salesforce and HubSpot for two decades is unraveling. The center of gravity in enterprise software is shifting to intelligent systems, and traditional CRM platforms are losing their gatekeeper status, being demoted to backend APIs for AI agents.
The article argues that the old moat of enterprise software was the high switching cost of accumulated data. But in the AI era, agents don't need traditional UIs — they can simultaneously pull structured data across calendars, email, Slack, and billing systems to make decisions. So the moat shifts from data accumulation to orchestration. Think of Facebook evolving from a friend graph to a news feed: the graph still exists, but the algorithm that recommends content now controls the user and the profits.
This technological shift brings two counterintuitive outcomes. First, AI hasn't reduced sales team headcount budgets. With AI-assisted selling, deal-close rates are significantly higher, and the high ROI actually encourages companies to invest more in people. Second, CRM call volumes have surged. That's because AI listens to calls in the background and writes structured notes back into the system, turning previously dormant databases into rich, active ones.
Despite the article's bullishness on the emerging intelligent orchestration layer, a16z's internal discussions suggest the power transfer is far more complex. Pure orchestration ability can easily be flattened by foundation model iteration. Traditional SaaS companies, stripped of their graphical interfaces, are unlikely to settle for being low-value pipes. Instead, they're well-positioned to evolve into permission-and-identity hubs governing all agent operations, leveraging years of accumulated business logic.