Steph Zhang: CRM Shifts to System of Intelligence
a16z GP Steph Zhang pointed out that enterprise software is shifting from "System of Record" (CRM database) to "System of Intelligence" (AI intelligent systems). CRM will not disappear but will become one of the input layers.
Similar to how the Facebook friend graph was replaced by the News Feed algorithm, AI agents now automatically complete tasks such as researching 10-Ks, real-time objection coaching, and call note rewriting, becoming the primary "intelligent feed" used by salespeople daily.
Market mechanisms show that GTM budgets are accelerating from traditional CRM seat licenses to AI orchestration and agent platforms, with funding flowing towards the intelligent layer built on Salesforce/HubSpot APIs. This trend drives capital towards AI-native GTM startups, putting pressure on pure record-based SaaS.
Source: Public Information
ABAB AI Insight
Steph Zhang has been tracking the impact of AI on sales processes at a16z. This System of Intelligence framework continues her observation of tools like Notion and Figma transitioning from data storage to intelligent workflows, emphasizing that AI will elevate human execution to supervision and decision-making.
In terms of capital pathways, a16z is shifting resources from traditional CRM integrations to multi-system context integration and compliance development by investing in AI agent and orchestration startups. The motivation is to capture the new TAM as GTM software expands from 5-10% budget share to significantly lower labor costs while enhancing ROI.
Similar cases include how News Feed allowed Facebook to transition from social graphs to algorithmic content distribution, and Salesforce AppExchange's transformation towards AI Agent Fabric. The GTM software industry is currently undergoing a transition from database dominance to intelligent orchestration control.
Essentially, this is a technological replacement: traditional CRM, as the core of value, is being replaced by the AI orchestration layer. The fundamental mechanism is that AI agents can simultaneously pull data from multiple sources like CRM, calendars, emails, and Slack, synthesizing actions in real-time, which humans cannot process in parallel. This leads to a shift in switching costs from "data migration" to "workflow and institutional memory migration," thus achieving a structural transfer of pricing power in enterprise software from data accumulation to intelligent leverage.
ABAB News · Cognitive Law
Databases will always be the foundation, but value will always run to the layer that can think.
When switching costs move from data transfer to workflows, the old kings become the new infrastructure.
The more AI agents there are, the more record systems resemble APIs; the more stable the backend, the more valuable the frontend.