Are All AI Startups the Same? Not Even Close.
The answer is no—and understanding the differences is key to identifying where the real value lies in today's AI ecosystem.
Before classifying AI startups, it's important to understand AI infrastructure: the underlying layer of technologies that makes artificial intelligence possible. This includes vector databases, MLOps platforms, APIs, and other tools used to train, deploy, and operate AI models. While these technologies are not end products themselves, they provide the foundation on which everything else is built.
From there, AI companies generally create value in three main ways.
AI-Powered SaaS
Cloud-based software applications that automate tasks within a specific industry or business function. The more vertical and specialised the software, the harder it is to replace. This category includes Vertical AI, which delivers industry-specific solutions; AI Applications, designed to solve individual use cases; and Embedded AI, where AI capabilities are integrated into existing software products, such as a CRM that automatically drafts emails.
AI Agents
Unlike traditional software, AI agents do more than provide recommendations or generate responses—they autonomously execute tasks, acting as digital employees. A key segment within this category is Workflow AI, where agents automate critical operational processes such as employee onboarding, account reconciliation, contract management, and other business workflows.
Service as a Product
Traditional services such as consulting, research, or analysis delivered through coordinated networks of AI agents. This model enables companies to scale operations while achieving software-like margins without sacrificing the level of customisation clients expect.
The Real Competitive Advantage: Data
So what determines whether an AI startup can build a lasting competitive advantage?
The answer is proprietary data.
The more unique data a company generates—and the better it uses that data to train and improve its models—the stronger its data moat becomes.
Building software has never been easier, and technology alone is no longer enough to stand out. The companies that will win are those solving real business problems and becoming deeply embedded in their customers' daily operations.
You can read the full newsletter in Encomenda Insights XIII on LinkedIn and watch Mario Brassesco's full video here:
cloud technology axon
Cuatrecasas has advised on the €15.1 million financing granted by pr...
