Incumbents like SAP are hard to displace because their value lies in the deeply embedded, customized business logic that defines a company's operations. Simply offering a database and APIs is insufficient, as it misses this crucial layer of operational DNA which acts as a key differentiator for the customer.
Productivity gains from AI don't simply reduce the total amount of work. Instead, they unlock new capabilities and analytical depths, creating new types of jobs and expanding what's possible. The long tail of work doesn't get shorter; it gets longer in a different, more complex way, representing a growing pie of innovation.
The core challenge for enterprise automation is not the 80% of standard workflows, but the 20% of exceptions. Almost everything interesting, from sales negotiations to customer service, is an exception. This is where human expertise and business differentiation lie, and it's the root of the challenge for AI agents.
During a tech shift like AI, the biggest opportunity for startups isn't direct competition. It's identifying the space between two established players who are cautiously bolting AI onto legacy products. This "in-between" space allows a startup to define a new category without being benchmarked against a 20-year-old feature set.
Enterprise software companies are not incentivized to offer true "headless" access via APIs. Doing so would turn them into a commoditized "dumb database," allowing others to capture the value layer. They will actively make it difficult to operate without their native interface to protect their business.
The term "agent" is largely a rebrand for programs that take a long time to run. In an enterprise context, their functions are best categorized as looking up data (easy), taking action (raises credential issues), or analyzing data (prone to hallucination). This framework helps demystify the current state of agentic AI.
Unlike consumer apps, enterprise software's most powerful network effect is internal. When a tool bridges the communication and workflow gap between previously siloed functions, like design and product development, it becomes embedded in the company’s collaborative fabric and is extremely difficult to remove.
The most "sticky" software is involved in core financial flows (like Stripe) or codifies complex regulations (like insurance or tax software). These systems are incredibly difficult to displace because they are tied to external forces like regulatory bodies, and the financial or legal risk of switching is too high for the customer.
Salesforce's announcement of a "headless" product was primarily a marketing move to acknowledge the rise of AI agents. In reality, it was a rebranding of their existing APIs, with no substantial change to the underlying product. This highlights a trend of incumbents using new jargon to appear innovative.