Are You Buying Efficiency, Transformation, or Strategic Optionality?

AI investment is often treated as one portfolio, creating confusion about value, governance, and returns. This episode explains why leaders need to distinguish efficiency, transformation, and strategic optionality before judging AI performance. 

It explores how different AI bets create value across different time horizons.

TLDR / At a Glance

• Efficiency gains in existing work
 • Transformation through workflow redesign
 • Optionality as future strategic flexibility
 • Category errors in AI business cases
 • Portfolio metrics by investment type
 • Governance matched to value logic

The key takeaway is that strong AI portfolios classify investments first, then apply the right proof standard for each category.

Transcript

The AI Spend Category Error

0:00

Are you buying efficiency transformation or strategic optionality? Most leadership teams are still treating AI as a single investment class. That is becoming a serious strategic mistake. This article explores why AI investments should be separated into three distinct categories efficiency, transformation, and strategic optionality. Each category follows a different economic logic, requires different governance, and should be measured against different expectations. The organizations getting this wrong are not necessarily underinvesting in AI. Many are misclassifying what they are actually buying. Why most AI business cases are built on a category error. Organizations are investing heavily in AI, but the returns remain uneven. PWC's 2026 AI Performance Study found that 74% of AI's economic value is being captured by just 20% of organizations. The problem is not simply whether organizations are spending too much or too little, it is whether leaders understand what kind of strategic bet they are making. Some AI investments are designed to improve the economics of existing work, faster processing, lower cost per task, higher throughput, or fewer manual steps. Others are intended to redesign workflows, operating models, or decision structures. Others are capability bets, investments in data foundations, workforce readiness, architecture, reusable AI components, or future strategic flexibility. These investments create value in different ways, over different timescales, and with different levels of uncertainty. Yet many organizations still evaluate them through a single AI business case lens. The same ROI logic is applied to cost reduction, operating model redesign, and long-term capability building. That category error creates distorted expectations long before any return is measured.

Efficiency Gains That Mislead

2:06

Efficiency is not transformation. Efficiency investments are the easiest category to understand because they improve the economics of existing work without fundamentally redesigning the operating model. Typical examples include coding assistance, customer support augmentation, invoice extraction, summarization tools, compliance drafting support, internal search and workflow triage. These investments usually operate within a bounded workflow and produce relatively visible proof points, lower task time, faster throughput, reduced cost per unit, or improved consistency. The customer support field experiment published in the Quarterly Journal of Economics is a useful example. Access to generative AI increased productivity by 15% on average and by roughly 30% for lower skill workers. Those are meaningful gains, but they exist inside a relatively constrained workflow. They do not automatically imply enterprise transformation. This distinction matters because efficiency gains are easier to prove and easier to overspend on. Cost reduction, throughput, and hours saved are highly legible to finance teams and boards. As a result, organizations can become trapped in a local optimization cycle where they automate narrow tasks while leaving the deeper architecture of work untouched. Efficiency is important, it builds credibility and creates early momentum, but faster tasks alone rarely create durable strategic advantage. Transformation is not a pilot scaled up.

Transformation Means Workflow Redesign

3:45

Transformation investments operate differently because they redesign workflows, operating models, and decision structures rather than simply improving isolated tasks. Transformation changes the operating system of the business, not just the efficiency of individual tasks. These investments are typically cross-functional, slower to mature, and much more dependent on leadership alignment, governance, and change management. They may involve redesigning end-to-end customer journeys, rebuilding planning systems, restructuring workforce allocation, or embedding AI into integrated operational flows. The strongest evidence consistently points in this direction. McKinsey found that workflow redesign had one of the strongest relationships with enterprise EBIT impact from generative AI. The World Economic Forum argues that AI creates compounding value when organizations redesign processes end to end, rather than layering tools onto fragmented workflows. MIT Sloan's workflow research similarly suggests that the biggest gains emerge from changing how tasks are sequenced, handed off, and coordinated, not merely from automating individual activities. Johnson ⁇ Johnson provides a useful example. Reuters reported that the company used AI to have the time needed to generate drug development leads and dramatically reduce clinical trial report preparation time. That is not simply a productivity tool layered onto existing work. It reflects redesign across connected research, trial, and operational workflows. JLL's AI-enabled software development work shows the same pattern. The World Economic Forum reported significant reductions in development effort after redesigning workflows across requirements gathering, testing, and development orchestration. The gains came not from one isolated AI feature, but from redesigning the broader flow of work. Transformation therefore requires a different leadership mindset. It should be governed as an enterprise change program rather than a technology deployment. The measurement model must also change. Leaders need to focus on cycle time, handoff reduction, decision latency, customer outcomes, margin improvement, and adoption of redesigned work patterns rather than only on task productivity.

Strategic Optionality And Future Freedom

6:13

Optionality is not waste. Strategic optionality investments are the hardest category to evaluate because their value is indirect and often delayed. These investments are designed to create future strategic flexibility rather than immediate operating gains. Examples include enterprise data foundations, AI architecture modernization, workforce capability building, internal experimentation platforms, reusable AI components, strategic partnerships, proprietary data assets, and large-scale AI upskilling programs. The point is not immediate payback, the point is preserving future strategic freedom under uncertainty. Optionality matters most when the future operating environment is still unstable. This category aligns closely with real options logic and discovery-driven planning. In uncertain environments, organizations sometimes need to invest before they fully understand where future value will emerge. The mistake is not making those investments, the mistake is pretending they should behave like short-term efficiency projects. Johnson Johnson's AI-driven skills inference initiative is a useful example. The company used AI to infer proficiency across future ready skills using learning and operational data. The value is not immediate cost reduction, it is future workforce visibility, capability allocation, and strategic flexibility. Accenture's enterprise-wide copilot rollout also partly fits this category. The immediate efficiency signals matter, but the deeper value lies in enterprise capability building, operating readiness, and long-term workforce adaptation. Optionality becomes dangerous only when it loses discipline. Without learning milestones, explicit hypotheses, staged funding, and clear review points, optionality spending can collapse into innovation theater.

Five Common Portfolio Mistakes

8:14

Why organizations get this wrong? Most organizations make five recurring mistakes. First, they fund by tool category rather than by economic logic. Completely different investments get grouped into a single AI budget, even though their value profiles are fundamentally different. Second, visible efficiency wins crowd-out transformation. Near-term savings are easier to explain than workflow redesign, future talent mobility, or architectural readiness. That creates a structural bias toward local optimization. Third, leadership pressure creates premature scaling. IBM found that 64% of CEOs invest before they fully understand the value because they fear falling behind. That pressure encourages organizations to scale shallow wins before redesigning the surrounding operating system. Fourth, pilot success is often misinterpreted. A successful bounded use case does not prove enterprise readiness. MIT Sloan's last mile problem captures this well. Models work, but value stalls because organizations fail to redesign workflows, governance, and operating structures around them. Fifth, boards and finance teams frequently demand the wrong proof at the wrong stage. Efficiency should prove near-term economics. Transformation should prove workflow and operating model change. Optionality should prove learning and capability creation. When organizations apply one ROI template to all three, they distort capital allocation.

Why One ROI Model Fails

9:56

Why one ROI model breaks the whole portfolio? The deeper issue is that these categories create value in different ways. Efficiency creates value by improving the economics of current work. Transformation creates value by redesigning how the business operates. Optionality creates value by preserving future strategic freedom. One measurement model cannot evaluate all three effectively. Efficiency should use direct operational metrics such as throughput, cost per task, error rates, ramp time, and realized savings. Transformation should use workflow and business metrics such as cycle time, service levels, handoff reduction, customer outcomes, decision latency, and margin impact. Optionality should use learning and capability metrics such as reuse rates, experiment velocity, talent readiness, architecture coverage, and strategic flexibility. This is where many AI portfolios break down. Leaders either force short-term ROI logic onto long horizon capability bets, or allow weak pilots to survive indefinitely because they are labeled strategic.

How Strong Portfolios Balance Bets

11:09

What strong AI portfolios do differently. A disciplined AI portfolio balances all three categories. Near-term efficiency investments build credibility, free capacity, and expose weak operational baselines. Transformation investments convert isolated gains into structural operating leverage through workflow redesign and process integration. Optionality investments preserve future strategic freedom while the competitive and technological landscape continues to evolve. The balance will vary by sector, maturity, data quality, and regulatory environment. But the important point is that the balance should be intentional. Strong portfolios also sequence correctly. A common pattern is to begin with targeted efficiency wins, then redesign high-value workflows, while protecting a smaller but deliberate pool of optionality investments in parallel. This is also where governance becomes critical. Efficiency initiatives can often sit close to business operations. Transformation requires enterprise level ownership and change governance. Optionality requires portfolio-style oversight with staged funding and explicit learning milestones. Across all three, responsible AI governance remains foundational. Governance is no longer separate from value creation. It is part of how value scales safely and sustainably.

Set Proof Standards By Category

12:37

The leadership decision. The practical issue is not whether AI has ROI. The issue is whether leaders are asking for the right kind of proof for the type of investment being proposed. An efficiency initiative should not be approved without a clear near-term operating case. A transformation initiative should not be judged only on immediate savings if it is genuinely redesigning workflows, decision rights, and operating structure. An optionality investment should not be allowed to drift forever under the cover of being strategic. This is the discipline leaders need. Classify the investment first, then set the proof standard. Efficiency should prove measurable operational improvement. Transformation should prove that the business is working differently. Optionality should prove that the organization is learning, building capability, or preserving future strategic choices. Without that distinction, AI portfolios become confused. Good long-term bets are killed too early, weak pilots are kept alive too long, and efficiency projects are mistaken for transformation.

Conclusion And Where To Read More

13:43

Conclusion. The strategic mistake is not underestimating AI. It is misclassifying AI spend. Leaders are not buying one thing. They are buying immediate efficiency, operating model transformation, and future strategic freedom. Treating those categories as if they obey the same return logic creates distorted expectations, poor governance, and weak capital allocation. The organizations that outperform in the AI era will not simply spend more intelligently, they will understand which investments change efficiency, which change the operating model, and which preserve the strategic freedom to compete in a rapidly changing market. This concludes the article. You can also read this article on my LinkedIn page where I share regular insights on AI, strategy, and emerging technologies.