How organizations can reskill and upskill employees in AI

12 steps to follow for reskilling employees in AI

Here are some steps that companies should follow when educating employees about AI.

1.      Start with a skills audit

Creating a current, comprehensive view of workforce capabilities is vital, including noting existing capabilities, critical gaps and role evolutions.

Carrying this out requires more than an HR spreadsheet. Use structured skills taxonomies and internal performance data to map current competencies across functions and use industry trends and future-of-work forecasts to identify priority skill areas. Scenario planning and skill adjacencies can help define how current roles can transition.

This audit should be revisited every two to three years to ensure it continues to reflect evolving technologies and job requirements.

2.      Develop a skilling roadmap

Once the gaps are clear, defining how the organization will close them is a good next step.

A one-size-fits-all training program won't work. Segment the workforce by exposure to AI, readiness for change and the nature of their roles. Offer foundational AI literacy to the full workforce and deeper, job-specific upskilling to groups in IT, products, marketing and operations.

3.      Redesign roles and update job descriptions

AI adoption leads to new responsibilities and job categories. Roles like AI risk manager, AI governance lead and AI auditor are increasingly necessary to ensure proper management of ethics, compliance and operational integrity.

Meanwhile, existing roles will evolve. Product managers and marketers will need to be proficient in AI, which necessitates rewriting job descriptions and creating plans for those who want to take on newly defined roles.

4.      Unlearn before learning

Many employees will need to unlearn outdated methods and assumptions before new practices can take root.

AI-related learning programs can help employees identify obsolete skills, challenge legacy behaviors and make room for different ways of thinking.

5.      Offer tailored and relevant training

To be effective, training must be problem-centered, business-aligned and immediately applicable.

Accomplishing these goals requires integrating learning into the flow of work, including using scenario-based modules and business simulations where AI concepts are tested against live challenges. Learning should be multi-format, and include cohort-based and self-paced. User engagement data can provide insight into how to change content if needed.

6.      Use AI platforms to accelerate learning

AI can support learning, not just be the subject.

Learning platforms with AI capabilities can personalize content delivery, recommend learning paths and assess mastery through adaptive testing, and virtual tutors and chatbots enable on-demand coaching.

7.      Upskill HR and enlist the C-suite

HR must transform alongside the rest of the company.

Learning leaders should possess fluency in AI, workforce analytics and digital learning infrastructure, and learning development teams should be able to rapidly deploy, iterate and scale training programs.

In addition, members of the C-suite must go beyond passive sponsorship. Executives should participate in AI training, publicly demonstrate their new skills and tie them  to employee performance assessments and promotion. Leadership buy-in remains one of the most influential variables in successful workforce transformation.

8.      Nurture change management skills

When AI adoption collides with employee growth, resistance is inevitable. Employees might view AI as a threat to career stability and progression.

To address this, a change management focus is crucial. Improving managers' communication skills, creating safe forums for feedback and connecting reskilling efforts to personal growth opportunities can all help. Managing behavioral change with the same rigor as technology change is vital.

9.      Create a culture that values learning

Many training programs fail because learning is treated as a task, not a core value. Surface-level familiarity with AI tools doesn't drive transformation.

Organizations must prioritize curiosity, reflection and experimentation as part of their culture to build real capabilities. They must also publicly celebrate learning and embed knowledge-sharing into team rituals. Organizations should combine structured learning paths with hackathons, ideation contests and sprints where employees can test and apply new AI skills.

10.  Think like a university

Every company must become a modern corporate academy.

Organizations cannot treat education as something that happens before employment. Learning must be aligned with business needs, and training must be integrated into career path planning.

11.  Use multiple tools

No single provider or platform can meet all learning needs.

Companies should combine internal training on proprietary tools and processes with training that's conducted by external partners. Employees should pursue industry-standard certifications.

12.  Train your ecosystem, not just your employees

A company's resellers, distributors, agents and partners are extensions of the brand. As AI reshapes customer expectations, partners should possess the same AI fluency and capabilities as internal teams, including product knowledge, platform literacy and the ability to explain AI-driven features or experiences to customers.

Neglecting partner training could introduce inconsistency and confusion.

Employee AI skills to cultivate

Organizations need to move now, while AI adoption is still uneven and its affect unevenly distributed.

The following are some of the AI skills that are worth cultivating, with skills listed by specific role.

All employees

All workers should possess the following skills:

• Data literacy. Employees should be able to interpret, validate and apply data effectively in AI-supported decision-making.
• AI collaboration. Employees should be able to work effectively alongside AI and understand when human input is necessary.
• Prompt engineering. Employees should be able to create and test prompts for generative AI tools.

• Adaptability. Employees should be able to pivot if needed as technologies and workflows evolve.
• AI governance and ethics. Employees should understand the ethical and regulatory boundaries involved with AI.

Technical teams

Employees in technical roles should possess the following skills:

• Rapid prototyping. Employees should be able to quickly design and make changes to AI-enabled tools or experiments.
• AI frameworks and tools. Employees should be able to use platforms, frameworks and tools for model development.
• Hallucination reduction. Employees should be able to apply techniques like retrieval-augmented generation or RAG, prompt tuning and post-processing, which can help minimize factual errors.
• RAG design patterns. Employees should be able to implement retrieval-augmented generation so responses are grounded in reliable, real-time data.
• Fine-tuning of LLMs. Employees should be able to adapt models to specific domains by carrying out supervised learning and hyperparameter tuning.
• Context engineering. Employees should be able to improve LLM output quality by managing memory, inputting boundaries and adjusting context windows.
• Tool chaining. Employees should be able to build compound workflows by connecting different AI tools or APIs in sequence.
• AI agent protocols. Employees should be able to design and manage autonomous agents that have controlled decision-making capabilities.
• AI user interface optimization. Employees should be able to develop UIs that guide and constrain AI functionality appropriately for end users.
• Model interpretability and explainability. Employees should be able to understand and communicate how AI models make decisions. That comprehension is critical for trust and compliance.

Machine learning ops and engineering

Employees in machine learning ops and engineering should possess the following skills:

• AI experimentation and sandboxes. Employees should be able to set up and manage sandbox environments for safe, contained AI testing prior to scaling.
• LLM pricing mechanisms. Employees should understand token-based billing, latency trade-offs and vendor pricing strategies.
• LLM routing. Employees should be able to assign queries to the best model based on applicable factors such as cost, latency or use case.
• AI cost optimization. Employees should be able to reduce compute and API spend through batching, routing or model selection strategies.
• AI tools and development standardization. Employees should be able to create enterprise policies and standards for tools, processes, open source and procurement.
• LLM evaluations and benchmarks. Employees should be able to run internal benchmarks and standardized tests to assess model performance.
• AI maturity models and feedback loops. Employees should be able to build systems that incorporate user feedback, which can improve AI models over time.
• AI documentation and compliance. Employees should be able to create and maintain traceable records of model decisions, training data and governance practices.
• Data engineering for AI readiness. Employees should be able to structure, label and clean data to make it usable for AI applications.

Managers and leaders

These leadership skills will determine whether AI adoption stays tactical or becomes transformational.

• AI literacy for strategic decision-making. Managers and leaders must understand what AI can and cannot do, know the difference between automation, augmentation and transformation, and know what questions to ask about AI.
• Business-AI translation. Managers and leaders must be able to connect business challenges to AI capabilities, including potential AI use cases and communicate about expected outcomes.
• AI governance and risk oversight. Managers and leaders must be able to direct employees about how to use AI responsibly and understand key concepts about AI bias, compliance and reputational risk.
• Change management in AI contexts. Managers and leaders must be able to anticipate AI resistance and communicate clearly about the technology as AI redefines workflows and roles.
• Critical thinking about AI outputs. Managers and leaders must know when to trust AI and when to question AI recommendations, including how to test its findings.
• Evaluation of AI tools and vendors. Managers and leaders must be able to evaluate AI products and ask about data use, model updates, security, integration capabilities and total cost of ownership.
• Team enablement and upskilling. Managers and leaders must be able to identify skills gaps and create space for experimentation and capability development.
• Human-AI role design. Managers and leaders must be able to lead team role redesign, including deciding when AI can add value and when humans must still be involved.
• AI foresight and scenario planning. Managers and leaders must be able to use AI tools when creating long-term plans.

HR and L&D

Those responsible for talent, workforce strategy and learning systems should possess the following capabilities:

• AI literacy for workforce strategy. HR and learning and development (L&D) workers should understand AI capabilities and limitations when using it with org design, talent planning and company strategy.
• Skills taxonomy and workforce mapping. HR and L&D workers should be able to use AI to build and maintain skills frameworks and use AI to discover future job needs.
• Reskilling and redeployment pathways. HR and L&D workers should be able to use AI tools to design internal mobility and reskilling strategies and identify adjacencies and gaps.
• Learning path design for AI skills. HR and L&D workers should be able to curate and structure AI learning journeys and tailor the learning to business needs.
• Change management for AI adoption. HR and L&D workers should be able to develop strategies to help employees deal with AI-related changes, including creating communication about changes, training and manager toolkits.
• AI-powered personalization of learning. HR and L&D workers should be able to use AI tools to create adaptive, role-relevant learning experiences at scale.
• Responsible AI in talent processes. HR and L&D workers should be able to assess whether AI tools that the company is using for hiring, performance, learning and workforce analytics meet standards for bias, explainability and compliance.
• Evaluation of AI learning vendors and tools. HR and L&D workers should be able to assess AI-powered learning platforms and content providers based on adaptability, integration, security and pedagogical value.
• Data literacy for people analytics. HR and L&D workers should be able to understand and act on workforce data.
• AI tool use in L&D operations. HR and L&D workers should be able to use AI to automate content creation, cohort management, learning assessments and knowledge curation.

Kashyap Kompella, founder of RPA2AI Research, is an AI industry analyst and advisor to leading companies across the U.S., Europe and the Asia-Pacific region. Kashyap is the co-author of three books: Practical Artificial Intelligence, Artificial Intelligence for Lawyers and AI Governance and Regulation.