The economics of knowledge work are changing in ways that are easy to underestimate. Before AI, people understood that producing competent work at scale required time, specialized skills, and organizational infrastructure. This meant that work activities like writing articles or building functional software came with meaningful costs. LLMs have driven those costs down to practically zero. Now AI’s output is the new average and “competent” execution is abundant.
When competent work is available at nearly zero cost, it provides little competitive advantage. We have seen versions of this value shift before: Cloud computing changed the skills required to manage infrastructure and GPS reduced the value of navigational skills. AI brought on a similar reset across a broader range of knowledge work.
For CEOs, this raises an important question: When knowledge work becomes abundant and cheap, where does the value move? My industry, search marketing, offers an early look at the answer. Search historically required considerable execution across content production, keyword research, technical recommendations, reporting, and analysis. Those capabilities needed specialized people and processes, along with technology. Executing this consistently and efficiently was a competitive advantage. AI lowered the cost of many of those functionalities while enabling a flood of new content and dashboards, as well as trackers and recommendations.
THE COMPETITIVE ADVANTAGE HAS SHIFTED
The barrier to production has fallen, but knowing which work produces meaningful outcomes remains a human skill. In my industry, AI could theoretically perform every step of a content workflow, from research and writing through quality review and publishing. But we need people throughout that process because expert decisions at each step affect the final outcome. People evaluate sources. They provide industry context. And they determine whether the activity is still aligned with the client’s goals.
This becomes particularly important during optimization. Businesses have always been susceptible to optimizing measurable proxies because they are easier to track than business impact. Search provides plenty of examples. Rankings, clicks, backlinks, traffic, and content volume have all been valuable measures because they correlated with outcomes that businesses cared about.
Those relationships, however, are changing. A ranking may produce less traffic as consumers increasingly get answers directly from search engines and AI platforms. Traffic may tell an incomplete story about whether a brand is visible during a customer’s research process. Content volume can increase dramatically without creating greater authority, awareness, or demand. Leaders, therefore, have to keep reassessing whether the SEO metrics their organizations optimize still correlate to the desired outcomes. Some might be tempted to throw the baby out with the bath water, i.e., abandon SEO.
BEWARE OF OPTIMIZING THE WRONG THING
Lower execution costs increase the price of bad decisions (see above). When organizations can produce content, software, analyses, and campaigns at unprecedented speed, flawed assumptions spread further before they’re challenged. The bottleneck shifts from producing work to deciding what work is worth producing.
That is where I expect experienced practitioners to continue creating significant value. Expertise includes the accumulated context that allows someone to recognize a weak proxy, question an assumption, connect information from different parts of a business, and identify when a seemingly efficient course of action is taking the organization in the wrong direction.
It also includes knowing when to stop. In an environment where almost anything can be optimized or automated, deciding what deserves neither may become one of the most valuable skills a person can develop.
BUILD ORGANIZATIONS AROUND JUDGMENT
This should make CEOs think about talent differently. Asking how much of someone’s existing job AI can perform gives leaders only part of the picture. They also need to consider how that person can use AI as a force multiplier. Employees who understand their field deeply have an opportunity to operate across a broader set of problems, applying their experience where judgment and context matter most.
Over the next several years, the tools will continue improving and many of today’s AI skills will become commonplace. Interfaces will become easier and models will gain access to more context. Tasks that currently require careful prompting will require less effort. That progression will continue pushing differentiation toward the expertise, experience, and perspective that people bring to the tools.
Every technological shift changes what is scarce, and scarcity is where economic value tends to accumulate. AI has already made competent execution abundant across a growing number of fields. The leadership challenge now is figuring out what deserves execution, then building organizations where people spend more of their time making those decisions well.
Patrick Briggs is CEO of Semify.