Rohit Kedia is a software architect turned CEO of Xoriant, a AI-native Applied Intelligence digital engineering company.
getty
Have you ever paused to think if AI is a wave to ride or a storm to survive? If you are a CIO or CTO at a banking, manufacturing, FMCG or high-tech company right now, you are almost certainly feeling both simultaneously.
Let me address the seemingly bad news first. The storm is real. Over $1.5 trillion has already been wiped from SaaS valuations. Enterprise software categories that looked permanent three years ago are being disrupted overnight. The tools your teams rely on are being repriced, consolidated or made redundant by AI at a pace that no technology planning cycle was built to handle.
The wave is equally real. Global AI solutions and services spending is projected to reach $1.3 trillion by 2029, growing at 31.9% annually. Eighty-eight percent of enterprises are now actively investing in AI across business functions. The competitive advantages being built right now by AI-forward companies are real, measurable and compounding.
But here is the problem. Most enterprise IT leaders are placing their AI bets in the wrong place.
According to McKinsey's State of AI 2025, only one-third of companies have begun to scale their AI programs at enterprise level despite 88% reporting active investment. The remaining two-thirds are spending—just not where it counts. Their budgets are concentrated at the model layer: platform licensing, infrastructure, vendor partnerships and proof-of-concept development.
The opportunity that is being left on the table sits entirely above that layer. This is the "apply AI" economy, which is the work of taking models that already exist and embedding them into real business operations (for example, into the retail supply chain that still cannot predict demand at a store level, or into the wealth management platform that still cannot personalize advice at scale).
Goldman Sachs makes a clarifying point in its March 2026 "Top of Mind" research report that enterprise leaders often miss in the noise: AI is software. Rather than eating the software market, AI is expanding it by lowering the cost of code creation while simultaneously raising the ceiling of what software can do. This is precisely why the "apply AI" economy is growing faster than most enterprise IT budgets are tracking.
So, the question is not whether to invest in AI, but whether your investment is going toward building the foundation that captures the expanding market.
There is a pattern I am seeing repeat itself across boardrooms right now. Every week brings another AI partnership announcement, another agent-to-agent demo, another bold claim about products shipped and workflows transformed. The commentary is loud and the theater is convincing.
And yet, when you look past the noise and ask what has actually changed in how value is created for customers, the honest answer is: not enough.
This is often because the four foundations that determine whether any of that AI investment actually delivers value remain chronically underfunded:
• Process: AI applied to a broken workflow produces broken outcomes faster. Every dollar spent on the model without redesigning the process around it is a dollar that produces activity, not outcomes.
• Technology: Fragmented legacy systems cannot support embedded intelligence at the point of decision. The architecture must be modernized to carry the load. Buying a more powerful model does not solve an integration problem.
• Skills: There is a critical difference between a workforce that knows what AI is and one that knows how to work with it. Yet, skills transformation remains one of the most consistently underbudgeted line items in enterprise AI programs, and it's often treated as a change management afterthought rather than a core delivery requirement.
• Data: No model, however capable, produces reliable intelligence from unreliable data. And yet data readiness receives a fraction of the investment that model selection and platform procurement does.
Gartner's February 2025 research found that through 2026, organizations will abandon 60% of AI projects due to insufficient AI-ready data. That is not a technology failure. That is a budget allocation failure.
The distinction between “using” AI and “applying intelligence” is ultimately a budget question as much as a strategy question.
Using AI means allocating spend toward tools layered on top of existing workflows—a co-pilot here, a chatbot there, an analytics layer on top of a dashboard.
Applying intelligence means something categorically different. It means embedding AI natively end-to-end into how a business creates and delivers value across platforms, data, processes and people in service of outcomes that are specific and measurable.
This requires a fundamentally different budget architecture, which can be thought of in three layers. The first is foundation investment. This is the unsexy work that makes everything else possible: data readiness, process redesign, legacy modernization and workforce fluency programs.
The second layer is intelligence deployment. This is where AI gets embedded into the actual workflows, not as a stand-alone tool but as a native capability within the platform, the product or the customer journey.
The third layer is human and agentic orchestration. The enterprises that generate real value from this layer are the ones that have already invested in the foundation and the intelligence deployment that make the orchestration meaningful.
Applying intelligence means shifting budget from the visible to the foundational, from the demonstrable to the durable, from the model layer to the application layer. It also means telling your board that the AI investment that will matter most next year is the one that looks least exciting this year.
But this is precisely the bet worth making.
So, the question is not whether your organization has an AI budget. It is whether that budget is building the foundation for applied intelligence or funding a very expensive AI theater.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。