Is Your New AI Vendor Just Three Enterprise Consultants in a Trench Coat?

I spent some time at Founded in FoCo last week, and one conversation kept coming up: the AI gold rush and the rapid rise of consultants and startups pitching "game-changing" solutions. But the more I listened, the more I was reminded of another era—one that many of us in tech remember a little too well.

If you've ever worked in an enterprise environment, you probably remember the heyday of SAP and ERP consultants. They swept into businesses with promises of efficiency, digital transformation, and industry best practices. What they often left behind were highly complex, deeply embedded systems that no one fully understood except, of course, them. Need a change? A report tweaked? A workflow adjusted? Better call your consultant—because everything was built in a way that made it practically impossible to modify without their help.

Fast forward to today, and I’m seeing AI vendors begin to tread the same well-worn path. They position themselves as essential architects of business-critical automation, but often, what they’re really selling is dependency. Instead of clean, explainable solutions, they offer black-box models with just enough obfuscation to keep clients locked in.

How to Spot an AI Vendor Selling Dependency, Not Solutions

If you're evaluating an AI-powered tool or consultant, ask yourself these questions:

• Do you know how it works? If the vendor can’t (or won’t) clearly explain how their solution processes your data, you’re handing over critical business processes to a black box.
• Is your data being used to train their models? Some AI vendors improve their product by feeding client data back into their training models. That might sound harmless—until you realize your proprietary insights could be strengthening a tool that your competitors also use. Always check the fine print on data usage policies.
• What infrastructure does it run on? If the AI solution is tightly coupled with a specific cloud provider or requires the vendor’s infrastructure to function, you might be stuck with them indefinitely. Ask whether the system can be deployed on your own infrastructure or migrated if needed.
• Is it built on open standards or proprietary tech? The more proprietary it is, the harder it will be to transition away if you need to. Custom models aren’t inherently bad, but they should be documented and maintainable.
• Who owns your data? If you can’t easily export your historical data and use it elsewhere, you’re not just buying software—you’re renting your own business intelligence.
• What happens if you stop using the service? Will your workflows grind to a halt? Will you need an expensive migration just to retain functionality? A good solution should be designed with portability in mind.
• Is the complexity necessary? AI solutions should simplify processes, not introduce unnecessary layers of abstraction that only the vendor understands. If you’re getting more confusion than clarity, that’s a red flag.

Automation Should Empower, Not Entrap

The goal of automation—whether AI-driven or not—should be to make your business more efficient, adaptable, and resilient. But too often, vendors use complexity as a strategy to keep you dependent, embedding themselves into your operations in ways that make switching costly and painful.

Before you commit to any automation solution, ask yourself: Does this make my business better, or just more reliant on an outside provider? True automation should streamline processes, provide transparency, and give you more control over your operations—not less.

If your vendor isn’t building for flexibility and long-term success, you’re not investing in automation. You’re buying a dependency.

Want to build automation that works for you—not for the vendor? Let’s talk.