AI Is Changing the Way We Think. Leadership Must Change the Way Organizations Decide.

About the Original Research

This Market Signals article responds to Boston Consulting Group's article, When Everyone Uses AI, Companies Risk Losing Critical Skills, authored by Sagar Goel, David Martin and Charikleia Kaffe.

The article explores an increasingly important question facing organizations adopting generative AI: if AI becomes deeply embedded in everyday work, what happens to the human capabilities that organizations have traditionally relied upon? Drawing on emerging research into cognitive offloading, organizational learning and workplace behavior, the authors suggest that while AI undoubtedly increases productivity, it may also reduce opportunities for employees to practice critical thinking, problem-solving and independent judgment if organizations become overly reliant on AI-generated outputs.

It is a thoughtful contribution because it shifts the conversation away from technology adoption and toward organizational capability. Rather than asking whether AI works, the article asks whether organizations are designing environments in which people continue to develop expertise alongside increasingly capable machines.

We believe that is precisely the right question.

The challenge facing leaders is no longer deciding whether AI belongs in the enterprise. That debate has largely been settled. The more important question is how AI should be integrated into organizations in ways that strengthen human capability rather than gradually replacing it.

As with every Market Signals article, our intention is not simply to summarize the original research, nor to challenge its conclusions. Instead, we examine what broader market signal it reveals, consider how it aligns with what we are seeing across industries and regions, and offer our perspective on where organizations should focus next.

In this case, BCG's warning about the potential erosion of critical thinking highlights a much larger opportunity. We believe AI's greatest contribution will not be writing more content or automating more tasks. Its greatest contribution will be helping organizations make better decisions through trusted information, stronger business context and more capable people.

Understanding the Question BCG Is Really Asking

One of the reasons Boston Consulting Group's article stands out is that it deliberately avoids the familiar debate about whether AI will replace jobs. That conversation has dominated headlines for the past two years, often reducing a complex organizational challenge to a simple question of automation versus employment. Instead, the authors ask something far more nuanced: what happens when AI begins participating in the thinking process itself?

That distinction matters.

For decades, enterprise technology has largely been designed to help people execute work more efficiently. Spreadsheets accelerated financial modeling. Customer relationship management systems organized sales activity. Business intelligence platforms made reporting more accessible. These technologies changed how work was performed, but they still depended on people to frame problems, interpret information and exercise judgment.

Generative AI represents a different category of technology. Rather than simply processing information faster, it can now draft reports, summarize research, generate recommendations, write software, analyze documents and respond conversationally to increasingly sophisticated questions. It is no longer just supporting cognitive work; it is participating in it.

BCG argues that this creates a new organizational challenge. If employees increasingly rely on AI to perform tasks that previously required reasoning, synthesis or critical evaluation, they may gradually lose opportunities to develop those capabilities themselves. Over time, organizations could become more productive while simultaneously becoming less capable. It is an uncomfortable proposition because the effects may not be immediately visible. Productivity metrics may improve long before leaders recognize that expertise, confidence and independent thinking have begun to erode.

The article draws on a growing body of research suggesting that people naturally offload cognitive effort whenever reliable alternatives become available. History offers many examples. Few people now memorize telephone numbers, calculate long mathematical equations by hand or navigate unfamiliar cities without digital assistance. These technologies made everyday life easier, but they also changed which skills people practiced regularly.

Generative AI extends this pattern into areas that have traditionally been considered uniquely human. Instead of recalling facts or performing calculations, AI can now generate arguments, structure ideas, identify patterns and recommend actions. The concern, therefore, is not that AI becomes intelligent. The concern is that people gradually become less engaged in the intellectual processes that develop expertise.

BCG's warning is not directed at the technology itself. Nor is it an argument against adopting AI. The authors acknowledge the enormous productivity gains AI can deliver and the competitive pressures organizations face to embrace it. Their argument is that leaders have a responsibility to ensure those gains do not come at the expense of the very capabilities organizations rely on to innovate, adapt and solve new problems.

That shifts responsibility away from the model and toward leadership. Technology does not determine organizational behavior. Leaders do. The way AI is introduced, governed and embedded into everyday work will influence whether it becomes a tool that strengthens human capability or one that gradually replaces it. The same technology can encourage curiosity in one organization and passive acceptance in another, depending entirely on the environment in which it is used.

We believe this is the most important contribution BCG makes to the conversation. The article asks leaders to stop evaluating AI solely through the lens of efficiency and start considering its long-term impact on organizational capability. That is a much bigger discussion than whether generative AI produces accurate answers. It is about how organizations continue to learn, how expertise is developed, and how judgment evolves in a world where intelligent systems increasingly participate in knowledge work.

At Data Tiles, we agree that these are the questions leaders should be asking. However, we also believe the discussion becomes even more interesting when we examine where critical thinking is changing rather than simply asking whether it is disappearing. The issue is not that AI thinks for us. The issue is whether organizations are intentionally designing environments where AI encourages people to ask better questions, challenge assumptions and make better decisions, or whether they are unintentionally creating cultures that reward accepting the first answer presented.

That, we believe, is the real market signal. It is not simply about protecting human judgment. It is about redesigning organizations so that AI expands human capability rather than quietly diminishing it.

The Signal We're Seeing: AI Isn't Reducing Critical Thinking. It's Changing Where Critical Thinking Happens.

Boston Consulting Group raises an important concern: if people increasingly rely on AI to perform cognitive tasks, they may gradually lose the opportunity to develop critical thinking, creativity and independent judgment. It is a compelling argument, and one that deserves serious attention from every executive considering how AI will shape the future of work.

At Data Tiles, however, we believe the signal emerging beneath that conclusion is even more significant.

Critical thinking is not disappearing. It is moving.

For centuries, expertise was often measured by what people could remember. Knowledge itself was scarce, so value came from possessing information that others did not. The internet fundamentally changed that equation by making information universally accessible. Today's competitive advantage is rarely having access to more information than everyone else. It is knowing which information matters, whether it can be trusted and how to apply it to a business decision.

Generative AI represents another shift in that evolution. Instead of simply helping us find information, AI helps us interpret it, organize it and explore it. That naturally changes where people apply their cognitive effort. Rather than spending hours producing a first draft or searching through multiple systems for answers, people increasingly spend their time evaluating outputs, refining questions, testing assumptions and determining whether an AI-generated recommendation makes sense within the context of their organization.

That is still critical thinking. In many ways, it is a more demanding form of critical thinking.

The quality of an AI response is directly influenced by the quality of the question being asked. Leaders who approach AI with vague objectives often receive vague answers. Those who understand the business problem, provide context and challenge assumptions consistently achieve better outcomes. The skill has shifted from generating every answer manually to framing better questions, recognizing weak reasoning, identifying missing context and knowing when experience should override an algorithm.

This distinction is often overlooked in discussions about AI de-skilling. History suggests that transformative technologies rarely eliminate human capability altogether. Instead, they change which capabilities become most valuable.

Calculators did not remove mathematics from education. They removed arithmetic as the primary barrier to exploring more advanced mathematical concepts. GPS did not eliminate navigation. It made travel accessible to millions of people while changing the value of memorizing every street and landmark. Search engines did not make expertise obsolete. They changed expertise from remembering information to evaluating and applying it effectively.

Artificial intelligence is following a similar trajectory. Used poorly, AI can absolutely encourage passive behavior. If organizations reward speed above all else, accept AI-generated outputs without scrutiny or discourage curiosity in favor of automation, BCG's concerns become very real. People may gradually outsource not just routine work, but independent thought itself.

Used well, however, AI has the potential to expand human capability rather than diminish it. It lowers barriers that have historically prevented more people from participating in complex analysis and decision-making. Employees who previously lacked access to specialist technical skills can now explore data, ask sophisticated questions and investigate scenarios that once required extensive analytical support. AI becomes less about replacing expertise and more about making expertise accessible to a broader group of people across the organization.

This is where our experience working with business leaders begins to diverge slightly from the prevailing narrative. The organizations we meet are rarely suffering from a shortage of intelligence. They employ experienced people who understand their customers, operations, products and markets exceptionally well. Their biggest frustration is not a lack of ideas. It is the inability to connect those ideas with trusted information quickly enough to make confident decisions.

The challenge is rarely thinking. The challenge is waiting.

• Waiting for data.
• Waiting for reports.
• Waiting for analysts.
• Waiting for engineering teams.
• Waiting for governance approvals.
• Waiting until the opportunity has already passed.

This is the market signal we believe leaders should be paying attention to. The greatest constraint on enterprise AI is not model capability. Nor is it human capability. It is decision latency, the time between a business question being asked and trusted information becoming available to answer it.

Every delay introduces uncertainty. Every additional hand-off increases friction. Every disconnected process slows the organization's ability to respond to customers, market conditions and emerging opportunities. AI may generate answers in seconds, but if trusted business information still takes weeks or months to assemble, the organization has not become meaningfully more capable.

This is why we believe the conversation needs to move beyond AI adoption and toward decision capability. The organizations that succeed will not simply deploy more AI. They will redesign the way trusted information flows through the business so that people can exercise better judgment while decisions still matter. AI becomes part of that process, but it is not the process itself.

Seen through this lens, BCG's warning is not simply about protecting critical thinking. It is an invitation for leaders to rethink how their organizations create, share and apply knowledge. The future will belong to organizations that use AI to strengthen human judgment, not by thinking instead of people, but by giving people faster access to the trusted business context they need to think more effectively.

The Market Signal: AI Doesn't Create Better Decisions. Better Decision Systems Do.

If there is one observation that consistently emerges from our conversations with executives around the world, it is this: organizations are not struggling because they lack artificial intelligence. They are struggling because they lack a reliable way of turning trusted information into timely decisions.

This is where we believe the market conversation is beginning to shift. For the past two years, AI adoption has largely been framed as a technology challenge. Organizations have focused on selecting foundation models, experimenting with copilots, investing in infrastructure and identifying use cases. Those investments are important, but they have also created the impression that enterprise value is primarily determined by the sophistication of the AI itself.

Increasingly, that assumption is proving incomplete. Recent research from McKinsey, Gartner, Forrester and others suggests that while AI adoption continues to accelerate, many organizations still struggle to translate experimentation into measurable business outcomes. The challenge is rarely that AI cannot generate an answer. It is that organizations have not redesigned the processes, governance and information flows surrounding those answers. Technology has advanced faster than organizational capability.

That is why we believe leadership must shift its attention from AI capability to decision capability. Decision capability is an organization's ability to move confidently from a business question to trusted evidence and then to informed action. It is not defined by how much data an organization owns or how many AI models it has deployed. It is defined by how consistently people can access trusted business context when important decisions need to be made.

Every organization already contains an extraordinary amount of expertise. Sales teams understand customers. Finance teams understand commercial performance. Operations teams understand supply chains. Marketing teams understand markets. Human resources teams understand workforce dynamics.

The problem is not that this knowledge does not exist. The problem is that it is fragmented across systems, reports, departments and governance processes that were never designed for the speed at which modern organizations are expected to operate.

As a result, organizations often create an unintended dependency on specialists. Business questions are handed to technical teams, data engineers, analysts or reporting functions, who spend valuable time locating data, validating definitions, reconciling inconsistencies and rebuilding information that already exists somewhere else in the enterprise. By the time trusted information returns to the business, the meeting has ended, priorities have shifted or the opportunity has passed.

Artificial intelligence does not solve that problem by itself. If AI is connected to fragmented, inconsistent or poorly governed information, it simply produces answers faster from information that may still be incomplete or misunderstood. The speed of the response increases, but the confidence behind the decision does not.

This is why we believe the organizations creating lasting value from AI are thinking differently. Rather than asking, "How do we use AI?", they are asking, "How do we improve the quality and speed of our decisions?"

That subtle change in perspective transforms the role AI plays inside the organization. Instead of becoming another destination for employees to visit, AI becomes one component of a much broader decision ecosystem. Trusted information, business context, governance, human expertise and artificial intelligence begin working together instead of existing as separate initiatives owned by different parts of the organization.

It also changes the way leaders think about people. Boston Consulting Group rightly cautions against allowing AI to replace critical thinking. We would take that idea one step further. The organizations most likely to succeed will be those that deliberately design environments where AI encourages better thinking.

• Where people ask more insightful questions because trusted information is immediately available.
• Where business experts spend less time searching for data and more time interpreting it.
• Where governance provides confidence rather than delay.
• Where AI becomes a catalyst for deeper conversations instead of a substitute for them.

This philosophy sits at the heart of what we describe as becoming a decision-driven organization. Decision-driven organizations recognize that competitive advantage no longer comes from simply collecting more data or deploying more AI. It comes from reducing the distance between a business question and a trusted answer. They understand that the true measure of AI success is not the volume of content it produces or the number of workflows it automates, but whether it enables people to make better decisions with greater confidence.

Ultimately, this is the market signal we believe leaders should be watching. The next generation of enterprise advantage will not belong to the organizations with the largest language models or the most AI pilots. It will belong to the organizations that redesign the way decisions are made, where trusted business context, embedded governance, human expertise and artificial intelligence come together to strengthen judgment rather than replace it.

That is where AI moves beyond productivity. That is where it begins creating genuine organizational capability.

Extending the Conversation: From AI Capability to Decision Capability

Boston Consulting Group's research identifies something important and, in our view, correct. When AI removes the effort involved in reasoning, it can quietly remove the opportunity for reasoning as well. We do not contradict that conclusion. We would like to extend it, because our work with enterprises suggests that the underlying cause sits one layer beneath the technology.

We do not believe AI is making people think less because AI is too powerful. We believe people lean on AI because trusted business knowledge is genuinely difficult to reach. Most organizations still operate using decision models designed for a pre-AI world, where a business question must travel through a long technical process, ticket queues, extract requests, definition debates, reconciliation and review, before trusted information becomes available to the person who asked it.

We describe the cost of that journey as decision latency: the elapsed time between a business question being asked and a trusted answer arriving in the hands of the person accountable for the decision. Decision latency is rarely measured, yet it shapes organizational behavior more than almost any other metric. When latency is high, people improvise. They rely on memory, on last quarter's spreadsheet, on intuition, and increasingly on whichever AI assistant answers fastest. Dependence on AI is not a failure of discipline. It is a rational response to a slow decision system.

This is why we believe the most useful shift leaders can make is from AI capability to decision capability, an organization's repeatable ability to move from a business question to trusted evidence to confident action. AI capability can be purchased. Decision capability has to be designed.

Our observation, and it is an observation rather than a prediction, is that the next source of competitive advantage will not come from larger models. It will come from redesigning the enterprise around trusted business context. That is where we believe the market goes next, and it is the conversation we want to help extend beyond the important problem BCG has named.

The Missing Layer: Business Context, Not More Data

Large language models understand language extraordinarily well. What they do not understand, and cannot infer reliably, is how a specific organization defines its own world.

• What actually counts as a customer, and when a prospect becomes one.
• How risk is classified, escalated and owned.
• Which margin definition the board uses, and how it differs from the one operations reports.
• When a shipment is genuinely late, versus late against an internal buffer.
• Which customers are priority customers, and who is allowed to decide that.
• Which assets are critical, and to whom.
• What a service level means in practice, including its exceptions.
• How the commercial hierarchy rolls up across regions, entities and channels.
• Which regulatory obligations constrain the answer before it is ever given.

None of these are technical facts. They are business knowledge. They live in the heads of experienced people, in policy documents, in exception handling, in the way a team has learned to interpret a number over many years. They are the reason two perfectly accurate reports can disagree, and the reason an AI assistant can be fluent and confidently wrong at the same time.

This is the missing ingredient between enterprise data and enterprise AI. Organizations have spent a decade making data available. Very few have made meaning available. Until business context is captured, governed and made reusable, every AI initiative is asking a model to guess what the business meant.

Business-Led Data Products: An Operating Model, Not an Architecture

Data products are often introduced as a technical architecture, a way of decomposing platforms into domains. We think that framing undersells them. In our view a business-led data product is something closer to institutional knowledge made durable.

• Reusable business knowledge — the definition is created once and trusted everywhere.
• Governed business context — meaning, policy, lineage and quality travel together.
• Organizational memory — expertise survives reorganizations, handovers and staff turnover.
• Trusted decision assets — designed for the decision they support, not the system they came from.

The critical word is business-led. When business experts package meaning alongside governance, lineage, quality and policy, the organization stops relitigating definitions and starts compounding knowledge. People and AI then operate from the same trusted understanding, which is the only durable way we know of to make AI answers and human answers agree.

Doing this well requires a particular kind of person: people who are fluent in the business and comfortable with data, able to translate commercial meaning into governed structure. We call them Purple People. They are, in our experience, the scarcest and most valuable capability in an AI-era enterprise, and building the operating model around them matters more than any individual technology choice.

Why Latttice Exists: The Business Activation Layer

Most organizations do not need another platform. They already own excellent ones. They have invested heavily in cloud data platforms, governance tooling, analytics estates and integration technology, and those investments were largely correct.

The unresolved challenge is different. It is enabling business experts to turn fragmented enterprise information into trusted, governed business knowledge that both people and AI can rely on. That is a capability gap, not a platform gap, and it sits between the systems that hold data and the people who hold meaning. We call it the business activation layer.

Latttice exists to operationalize that layer. Business users create trusted business-led data products directly. Governance is embedded rather than applied afterwards. Business context is captured at the moment of creation, by the people who actually hold it. The result is that AI no longer reasons over fragmented enterprise data. It reasons over trusted business knowledge.

Why Lenz Works: AI That Starts From Trusted Business Knowledge

Lenz is not effective because of the model behind it. Models are becoming a commodity and will continue to improve regardless of who builds them. Lenz is effective because of what it reasons over.

It reasons over trusted business-led data products rather than disconnected enterprise systems. It understands business context because the business has already created that context. It inherits governance because governance already lives inside the data products it uses. Nothing has to be re-explained to the AI, because nothing was left implicit in the first place.

That is why trustworthiness improves so noticeably. The AI is no longer inferring meaning from raw enterprise data and hoping it guessed the organization's definitions correctly. It begins from knowledge the business has already agreed on, governed and published. Trust becomes a property of the operating model rather than a claim about the model.