Paraphrasing William Gibson, the future of AI is here, but it’s nowhere close to evenly distributed yet. Last week in London, two conversations about enterprise AI demolished any neat narrative. In one meeting, the head of engineering at a large hedge fund described teams with fleets of agents in full production, and he personally relies on LLMs for all code writing—though junior hires are barred from using them. In another, a data engineer at a large retail bank painted the opposite picture: no agents and sparse LLM usage. His division, at least, is not moving fast on AI.

This divergence is not about one company “getting” AI and another not. It’s a reminder that even within the same organization, adoption curves can be wildly different. AI is widening the gap between teams that can absorb it operationally and those that cannot. Recent data supports this. McKinsey found that 88% of respondents use AI in at least one business function, but only about one-third have begun scaling AI programs. For agents, 23% report scaling an agentic AI system somewhere in the enterprise, while 39% are still experimenting. In any given function, no more than 10% say they’re scaling agents. Broad usage is not deep institutional change. There is still time to figure out AI; you are not behind.

Cue the engineering boom

The notion that “finance is cautious” or “regulated industries are behind” is too simplistic. Some financial firms are aggressive, some are not, and some teams inside the same firm are doing both. Deloitte’s 2026 enterprise AI research reinforces this: only 25% of respondents moved 40% or more of their AI pilots into production. Just 34% say they’re using AI to deeply transform their businesses—a number likely aspirational—while 37% use it at a surface level with little process change. This looks less like a tidal wave and more like a messy, uneven organizational test.

This explains why the “AI will wipe out software jobs” talk misses the point. The interesting thing about AI coding tools is not that they make software cheaper, but what companies do with that lower cost. Box CEO Aaron Levie invoked Jevons paradox: when a capability becomes cheaper, demand often rises. Cloud computing didn’t reduce compute needs; it encouraged building more things that consume compute. AI-assisted coding may do the same for software.

Engineering job data supports this. Lenny Rachitsky highlighted that engineering openings are at their highest in over three years. TrueUp data shows 67,665 open engineering jobs as of March 2026, up 78.2% from the recent low. Importantly, this is not concentrated at the top: 44.6% of posted roles are entry and mid-level, versus 38.3% senior and 13.8% senior-plus. AI is not eliminating roles for junior developers; it’s changing what enterprises want from engineers.

Stack Overflow’s 2025 survey found 84% of respondents using or planning to use AI tools in development, with over half of professional developers using them daily. McKinsey’s research shows the highest-performing AI-driven software organizations see 16% to 30% improvements in productivity, customer experience, and time to market, along with 31% to 45% improvements in software quality. But these gains come from reworking roles, workflows, and the entire product development system—not from sprinkling copilots over an unchanged process. That is a much harder organizational challenge than buying licenses.

Software engineering is alive and well

Returning to the London conversations: the hedge fund leader may be an early glimpse of where parts of enterprise engineering are headed. Less time hand-authoring code, more time specifying, reviewing, steering, and orchestrating systems that generate code. But the retail bank division is not irrationally lagging. In a heavily regulated environment, governance is the hard part. Deloitte reports that only 21% of surveyed companies have a mature governance model for autonomous agents (and those 21% are probably kidding themselves). Meanwhile, 73% cite data privacy and security as a top risk, and 46% cite governance capabilities and oversight. This is not bureaucracy for its own sake; it’s a recognition that plugging non-deterministic systems into deterministic, compliance-heavy environments gets messy fast.

Caution, however, is not free. Every quarter spent in pilot mode is a quarter in which more aggressive peers build operational muscle. OpenAI’s enterprise usage data shows how uneven that muscle-building already is. Frontier workers—the 95th percentile of adoption intensity—send six times more messages than the median worker. Frontier firms send twice as many messages per seat. OpenAI says the primary constraints are no longer model performance or tools, but organizational readiness and implementation. This rings true: the real divide is increasingly not between companies that have access to AI and those that don’t, but between teams that have learned to integrate AI into repeatable work and those still treating it as a promising but dangerous sideshow.

The distinction of task versus job matters. Writing boilerplate code is a task. Engineering is a job. Jobs bundle judgment, trade-offs, accountability, architecture, security, integration, testing, and the ugly reality of operating systems in the real world. AI can automate more tasks, but it hasn’t eliminated the need for jobs, especially in environments where bad software decisions carry operational or regulatory consequences. McKinsey’s broader AI survey found that most organizations are still navigating the transition from experimentation to scaled deployment, and high performers stand out because they redesign workflows and treat AI as a catalyst for innovation and growth, not just efficiency. That is very different from “we gave everyone a chatbot and now we need fewer people.”

So no, AI isn’t marching toward one uniform enterprise future where software engineers fade away. Instead, AI is splitting enterprises into fast-learning and slow-learning teams. It rewards organizations that redesign work, govern risk, and turn lower software costs into more software, not less. The code may be getting cheaper, but the ability to decide what should be built, how it should fit together, and how to keep it from breaking the business continues to increase in value. That’s not the death of software engineering—it’s the repricing of it, and every company and every team is paying different prices.

Source: InfoWorld News