Why Every Employee Will Need to Use AI in 2025Why Every Employee Will Need to Use AI in 2025

Over the past year, we’ve seen organizations differ in their approaches to AI. Some have taken every opportunity to embed AI in their workflows; others have been more cautious, experimenting with limited proof-of-concept projects before committing to larger investments.  

But unlike past technology breakthroughs that were only relevant for specific employees, AI is a horizontal skill. Business leaders need to embrace this fact: Every single employee needs to become an AI employee.  

In 2025 and beyond, we will start to see the difference between companies that treat AI as a feature and those that view it as a transformation. Here's how business and learning leaders should think about AI adoption throughout their organization.  

Establishing an AI-Ready Skills Vision 

For businesses to develop an AI-ready workforce, they need to establish a skills vision that sets out which employees require which level of competency. This vision shouldn't be permanent; instead, it should evolve in response to technological advances and the needs of the business. 

There are two ways of structuring an AI skills vision. The first is simple: builders and users. A small portion -- roughly 5% -- of an organization’s workforce will require the expertise to build AI systems, products, evaluation tools and language models. The remaining 95% simply need to know how to use AI to augment and accelerate their existing workflows.  

Related:Demand and Supply Issues May Impact AI in 2025

For a more detailed framework, leaders can break down their workforce into four levels: 
Center of excellence: Synonymous with “AI builders.” Think about data scientists, machine learning engineers, and software engineers. Their entire role is to design, build, and refine AI tools for internal or external clients. 

“AI + X”: These are the subject matter experts whose roles can be reimagined with the addition of AI. Employees at this level could come from a wide range of backgrounds, from mechanical engineers to finance leaders. AI can help these employees build something truly meaningful in their specific area of expertise. 

Fluency: At the fluency level, you don’t need to know how to use AI tools or apply them to your workflows. Instead, fluency is the required level for employees who are interacting with a technical counterpart. For example, a marketer selling a highly technical AI product needs a certain level of understanding to be able to accurately and effectively market that product. 

Literacy: This is the basic level of AI skills needed for front-line workers and individual contributors. AI literacy could help these employees boost productivity depending on their role and responsibilities. But it’s equally important for these employees to be part of the broader cultural change. A company is in a better position to innovate when every employee has achieved a standard level of AI literacy. 

Related:What Happens if AI No Longer Has Access to Good Data to Train On?

Avoiding Dangerous Amateurs 

For an organization to make the most out of AI, it needs to know the precise skill levels of its employees and where they need to grow in the future. 

For example, a company’s solutions will only ever be as good as their best contributors. Organizations must do everything they can to maximize the abilities of their Center of Excellence employees, because they set the bar for the rest of the organization. At one software company, I saw leaders transfer an expert in clean coding to a team struggling with code quality; improvements were evident across the organization within weeks, demonstrating the contagious nature of expertise. 

But, while experts should be placed at the forefront and driven to achieve more, organizations must be careful not to give the same opportunities to those who overstate their abilities. My friend and collaborator Fernando Lucini refers to these employees as “dangerous amateurs,” and they can slow down an organization’s progress with AI. As companies transition from prototyping to productizing an AI solution, they may realize that the experts they were counting on don’t have the skills needed to bring the product to market. Meanwhile, competitors with an accurate measure of employee skill levels will race ahead. 

Related:AI Risk Management: Is There an Easy Way?

Create the Foundation for Innovation 

For companies to innovate, they need to be able to adapt quickly to changing technologies and skills demands. In 2016, one of my most important tools was TensorFlow, a commonly used programming language. Less than a decade later, TensorFlow has evolved so much that I can no longer use it effectively without retraining and updating my skills. Highly technical skills perish quickly. 

Employees must establish a strong foundation in durable skills in order to master the perishable, cutting-edge technical skills. OpenAI built ChatGPT using innovative, breakthrough technologies. However, they could only create ChatGPT by drawing on their foundations in durable skills like mathematics, statistics, coding and English. AI-ready companies will need to embrace a T-shaped approach to skills development, combining a broad base of horizontal skills with a narrow set of deep, vertical skills. Innovation breaks through as a result of perishable skills but sustains as a result of durable skills. 

Every company is becoming an AI company. Every employee will need to use AI. Those who don’t embrace the change will inevitably fall behind.