Despite the hype, LinkedIn found that third-party AI tools for coding and design don't work out-of-the-box on their complex, legacy stack. Success requires deep customization, re-architecting internal platforms for AI reasoning, and working in "alpha mode" with vendors to adapt their tools.

For established firms like VCs, the primary challenge in adopting AI isn't change management or model selection. It's the painstaking process of migrating and cleaning decades of financial data from outdated systems to make it accessible and useful for modern AI agents.

Enterprises are trapped by decades of undocumented code. Rather than ripping and replacing, agentic AI can analyze and understand these complex systems. This enables redesign from the inside out and modernizes the core of the business, bridging the gap between business and IT.

AI is not a silver bullet for inefficient systems. Companies with poor data hygiene and significant technical debt find that implementing AI makes their bad systems worse, simply scaling the noise and dysfunction rather than solving underlying problems.

Enterprises are finding immediate, high return on investment by using AI to port legacy codebases (like COBOL) to modern languages. This mundane task offers a 2x speed-up over traditional methods, unlocking significant infrastructure savings and even driving new developer hiring.

Large enterprises operate on complex webs of legacy systems, compliance controls, and fragile integrations. Their high risk aversion and lengthy change management cycles create a powerful inertia that will significantly delay the replacement of established B2B software, regardless of how capable AI agents become. Enterprise architecture moves slower than market hype.

Counterintuitively, GitHub discovered that training coding models on more private enterprise codebases (e.g., modern web frameworks) provides little benefit. The significant performance gains come from training on scarce, legacy code like COBOL, where public data is limited but enterprise demand for modernization is high.

AI coding assistants have recently crossed a critical threshold. They are no longer just for building new features but are now highly effective at refactoring legacy code. This dramatically changes the economics of modernizing established software companies by accelerating the notoriously slow process of paying down technical debt.

AI can generate code, but the real value of enterprise software is its integration into complex human workflows, the massive costs of change management, and network effects. These human-centric problems create a durable moat that code generation alone cannot overcome.