Beyond the Hype: Where Enterprise AI Is Actually Delivering Value
The Data Is In: AI Usage Is Not What You Think
On a recent episode of AI Daily Brief, we highlighted a pivotal study from OpenRouter and Andreessen Horowitz (a16z) that analyzed an unprecedented 100 trillion tokens of real-world AI usage.[2] The findings paint a clear picture: the most impactful and economically significant applications of AI in the enterprise are not the ones grabbing headlines. Instead of broad, horizontal use cases, the data shows a surge in two primary areas: highly complex, vertical-specific programming and the rise of multi-step, agentic workflows. According to the study, programming has exploded from just 11% of usage at the start of 2025 to over 50% by year-end[1], while the use of reasoning models has become the new default, accounting for more than half of all tokens consumed. We are talking about feeding platforms an entire codebase to find a bug, a reality that fundamentally changes how we must think about AI strategy.
From Hype to Reality: The Strategic Implications for Your Business
The OpenRouter study, corroborated by recent findings from firms like McKinsey and BCG[3], exposes a crucial gap between the perception of AI and its practical application in the enterprise. While many organizations are still in the experimental phase, dabbling with chatbots and simple productivity tools, the market leaders are leveraging AI for deep, complex, and value-driven tasks. This creates a significant strategic challenge for executives. The data shows that demand for high-quality models is "wildly price inelastic," with users happily paying 10-50x more for a model that saves them ten minutes of debugging[1]. This indicates that value, not cost, is the primary driver of adoption for serious use cases. Furthermore, the market is not converging on a single "best" model; rather, a multi-model reality is emerging, where different models are used for different tasks. For example, the study found that Anthropic's Claude is used for over 80% of programming tasks[1], while other models dominate creative or role-playing scenarios. This fragmentation, combined with the fact that, according to McKinsey, only 1% of companies have achieved true AI maturity[3], underscores the complexity of deploying AI at scale. The challenge is no longer about whether to adopt AI, but how to navigate a fragmented, rapidly evolving landscape to drive real business value.
Key Questions for Your Leadership Team
The gap between AI hype and on-the-ground reality requires a candid internal assessment. The most prepared organizations are not just asking "what can AI do?" but are instead debating a more difficult set of questions. These are the conversations that separate the companies that will merely survive the AI transition from those that will lead it:
-
Are we measuring the right things? Given that value trumps cost, how are we evaluating AI initiatives beyond simple cost-saving metrics? Do we have the KPIs in place to measure the impact on complex workflows, developer productivity, and time-to-market?
-
Is our AI strategy aligned with real-world usage? Are we still focused on general-purpose chatbots, or are we identifying the high-value, specialized workflows where AI can deliver a step-change in performance?
-
How are we preparing for a multi-model world? Do we have a strategy to evaluate, select, and govern a portfolio of AI models, or are we placing a single bet on one provider?
-
Is our data and infrastructure ready for complex AI? The study shows a 400% increase in prompt size[1], with users inputting entire codebases. Is our data architecture prepared for this new paradigm of complex, context-heavy AI interactions?
-
Do we have the talent and governance to scale? With 74% of companies struggling to scale AI due to data governance issues[3] and a significant talent gap, what is our plan to build the internal capabilities required for enterprise-wide deployment?
How Leading Organizations Find Answers
The questions above can feel daunting, but they are not unanswerable. In our work with enterprise clients, we've observed a clear pattern: the organizations that successfully navigate AI transformation don't just buy technology; they build a strategic framework that aligns their people, processes, and data with their ambitions. They move from ad-hoc exploration to a structured, holistic assessment of their capabilities. They create a common language for the executive team to debate and prioritize. Most importantly, they get an objective, data-driven baseline of where they are today before they try to build for tomorrow. This is the philosophy we've codified into our AI Readiness Audit. It's not a product pitch; it's a diagnostic process designed to provide the objective clarity that leadership teams need to move forward with confidence. It allows organizations to systematically uncover their unique blockers and identify their highest-impact opportunities.
Continue the Conversation
Every organization's journey with AI is unique. If the questions and challenges discussed in this post resonate with your team, we welcome a conversation. We read all our emails and welcome conversation-ping me at danv@besuper.ai. Our goal is to help leaders build a clear, actionable roadmap. To learn more about how a structured assessment can de-risk your AI investment and accelerate your path to value, you can explore our approach at besuper.ai or reach out to our team to discuss your specific situation.
References
[1]OpenRouter State of AI Study - https://openrouter.ai/state-of-ai
[2]Andreessen Horowitz (a16z) State of AI - https://a16z.com/state-of-ai/
[3]McKinsey Global Survey on AI 2025 - https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
