Software in the Age of Agents | The a16z Show
TL;DR
The rise of AI agents is accelerating a shift toward 'headless' software where APIs and business logic matter more than user interfaces, but speakers warn that underestimating the complexity and stickiness of enterprise systems—where decades of regulatory compliance and institutional knowledge create deep moats—is a critical error as the industry moves beyond human-centric workflow design.
🔄 The Headless Transition 3 insights
Salesforce's headless announcement is largely marketing
Salesforce's 'headless 360' launch was primarily a rebranding of existing APIs to acknowledge the agentic shift, with no significant technical changes to how systems expose data.
Agents bypass traditional UIs entirely
Usage of Slackbot agents to access CRM data has increased 300%, demonstrating that agents interact with systems of record via APIs and chat interfaces rather than logging into graphical user interfaces.
Value migrates from workflow to data layer
In an agentic world, the stickiness shifts from UI and workflow capture to the underlying data, logic, and business rules stored beneath the interface.
🤖 Agent Capabilities and Constraints 3 insights
Agents perform three distinct functions
Enterprise agents generally execute lookups (simple data retrieval), actions (system changes requiring credentials and impersonation), and analysis (multi-system research that risks hallucination without verification).
Action capabilities introduce licensing complexity
When agents perform actions in systems of record, enterprises face unresolved questions about whether the agent constitutes an additional paid seat or falls under existing user credentials.
Current agents are often rebranded programs
Steven Sinofsky notes that many 'agents' are essentially long-running programs with new branding, and many headless APIs merely provide forgiving interfaces for lookups rather than transformative new capabilities.
🏢 The Reality of Software Stickiness 3 insights
Stickiness derives from invisible organizational entanglement
Enterprise software becomes sticky through UI muscle memory, undocumented standard operating procedures, downstream financial dependencies, and compliance requirements that embed vendors deeply into company operations.
SAP encodes business logic, not just data
Systems like SAP cannot be replaced by simply combining Postgres with APIs because they encapsulate decades of unique business rules, regulatory logic, and company-specific processes that constitute the actual enterprise.
The stickiest features are often accidental
The most powerful retention mechanisms, such as Outlook's calendar delegation capabilities, are frequently discovered only when customers threaten churn rather than being intentionally designed as lock-in features.
⚠️ Misconceptions About Enterprise Disruption 2 insights
Vibe coding cannot replace enterprise complexity
There is a 'wild underestimation' that modern AI tools can easily replace systems like SAP, ignoring the reality that enterprise software persists because it manages customization, tax laws across jurisdictions, and regulatory compliance that took generations to encode.
Legacy systems persist due to regulatory capture
Insurance and banking software written in COBOL decades ago remains unreplaceable because it codified external regulatory forces and compliance requirements that pure data layers cannot replicate.
Bottom Line
Companies building for the agentic era should focus on encapsulating complex business logic and regulatory compliance rather than just exposing raw data via APIs, as the true competitive moat lies in encoding institutional knowledge that cannot be easily migrated to new database layers.
More from a16z Podcast
View all
a16z Goes Global: Why American Tech Must Lead the World
a16z outlines its global expansion strategy to ensure American technological leadership in the AI era, arguing that AI models encode cultural values and that the U.S. must help allies adopt Western technology to prevent authoritarian 'digital colonization' and maintain national security through innovation.
Jake Paul on Going From YouTube to Boxing to Investing | a16z ft. Anti Fund
Jake Paul and Anti Fund co-founder Geoffrey Woo announce their oversubscribed $100 million growth fund backed by tier-one LPs like Andreessen Horowitz, while explaining how Paul's resilience and analytical operator mindset translate from entertainment to backing frontier tech companies.
The Media Game Has Changed
The media landscape has shifted from defense-oriented legacy outlets demanding plastic, controversy-free performances to an offense-driven ecosystem where authenticity and founder-led personal brands dominate; success now requires treating public conversations like private ones and building direct channels rather than relying on traditional press.
Why AI Feels Like the Internet in 1997 | Benedict Evans on a16z
Benedict Evans compares today's AI landscape to the internet in 1997, arguing that agentic coding has emerged as the first true product-market fit use case while the industry grapples with severe infrastructure scarcity and an uncertain future where foundation models risk becoming commoditized infrastructure rather than value-capturing platforms.