The Problem with Enterprise AI: My Take on the Three Biggest Challenges

I read an article based on the Stack Overflow 2025 Developer Survey that perfectly captured the core challenges we face with enterprise AI. The survey found that while 78.5% of developers are using AI tools, a significant 45.7% distrust AI-generated code. This highlights a critical disconnect.

The article identifies three key problems, but I have a different take on the best solutions.

Challenge 1: Executive Mandates The article points out that many leaders push for rapid AI adoption without understanding the risks, creating a divide between executive expectations and developer reality.

I agree. The real issue is a lack of strategy. Instead of mandating its use, we should empower developers to use AI as a tool to augment their skills, not replace them.

Challenge 2: Inconsistent Code Quality The article suggests that some companies require developers to submit prompts with pull requests to mitigate this. I find that approach reactive. To me, that is letting it do the wrong things first and then cleaning up the mess.

The solution is to give AI a spec to follow. The usage of AI should start earlier, in the planning stage. We can have AI generate the initial plan and file the Jira ticket. Then, when it helps with coding, it has a clear spec to follow, leading to higher quality results from the start.

Challenge 3: Lack of Internal Knowledge The article states that LLMs lack a company's proprietary knowledge, hindering their effectiveness.

I don't believe a company needs to build a massive knowledge base integration. A better solution is to start small with instructions files. Tools like Claude Code, Gemini CLI, and GitHub Copilot with VSCode all support this. These files can tell the AI about your coding style, the libraries to use, implementation strategies, best practices, and where to look for more information. These instructions can be checked into the repo for project-level guidance or shared among users.

The debate is not just about adopting AI, but about implementing it properly.

Article Link: https://thenewstack.io/solving-3-enterprise-ai-problems-developers-face/