Introducing AI Agents: The Next Evolution in Artificial Intelligence

Until recently, working with AI has been like a one-way street: you ask a question, it gives you an answer. But AI agents are more like a roundabout – they keep circulating through tasks until they achieve the desired result. These agents don’t just respond; they plan, act, verify, and adapt, often without human intervention at each step.

As learning professionals, we’re at a critical juncture. The emergence of AI agents represents more than just another technology trend – it’s a fundamental shift in how we can approach learning design and delivery using AI. Those who understand and harness this shift will be in a position to revolutionize workplace learning.

Let’s break this down with an example from training and development. When you ask a traditional Large Language Model (LLM) a question like “What are the best practices for safety training?” it will provide an answer based on its training data – essentially a literature review of known best practices.

An AI agent, however, would approach this task very differently. If you asked, “What safety training do we need for our warehouse team?” it would:

• Make a plan to gather relevant information
• Access your company’s incident reports and safety records
• Review current OSHA requirements for your industry
• Analyze employee skill assessments and training histories
• Compare current practices against regulatory requirements
• Identify specific gaps and needs
• Provide a customized recommendation based on actual data

This “loop” of actions makes agents fundamentally different from traditional LLMs. They don’t just provide generic information – they can gather specific data, analyze it, verify requirements, and create targeted recommendations – all without needing human input at each step.

The Building Blocks of AI Agents

Understanding these components isn’t just technical knowledge – it’s strategic insight that will help you make better decisions about which AI solutions to adopt and how to implement them effectively.

AI agents require five key components:

• Data inputs (their knowledge source)
• Models (the AI “brain”)
• Tools (their capability set)
• Interface (how users interact with them)
• AI expertise (the human design that pulls it all together)

Current State of Agent Development

While the potential of agents is tantalizing, they’re not quite “ready for prime time” yet. They’re promising enough to attract significant attention, but not reliable enough to be worth paying for yet. So far, two main approaches are emerging in the AI agent space, each with different implications for learning professionals:

1. Model-First companies (like OpenAI and Anthropic) are betting on creating more powerful AI models. Think of these as building better “brains” for AI agents. For learning professionals, these more powerful models could mean agents that better understand learning objectives, create more engaging content, and interact more naturally with learners.

2. Workflow Application companies are focusing on practical applications using existing models. These companies take existing AI capabilities and build specialized tools for specific tasks. For example, they might create agents that specialize in curriculum design, learning analytics, or personalized coaching.

The difference matters because it affects how we might use these tools. Model-first solutions might offer more sophisticated capabilities but require more setup and customization. Workflow applications might be more limited in scope but easier to implement for specific training needs.

What This Means for Learning Professionals

While some are still debating whether to use AI in the future, forward-thinking learning professionals are already identifying specific applications that deliver measurable value today. Here are key areas where AI agents could transform our work:

Content Development

• Automatically update training materials when policies or procedures change
• Create multiple versions of content for different learning styles
• Generate practice scenarios based on real workplace situations
• Develop assessment questions and feedback loops

Learner Support

• Provide 24/7 coaching through complex tasks
• Offer personalized learning paths based on performance data
• Create custom job aids on demand
• Monitor progress and suggest interventions
• Administrative Tasks
• Schedule and coordinate training sessions
• Track compliance requirements and certifications
• Generate progress reports and analytics
• Manage learning resource libraries

Performance Support

• Create just-in-time microlearning modules
• Assist with complex problem-solving in real time
• Provide contextual help during workflow
• Facilitate peer learning connections

Getting Started with AI Agents

As learning leaders, our role isn’t just to implement technology – it’s to guide our organizations through this transformation thoughtfully and effectively. Here’s how to begin:

Start Small

• Use existing tools like ChatGPT plugins to automate simple tasks
• Experiment with browser extensions that can help research and organize content
• Try task automation tools for routine administrative work

Build Your Understanding

• Follow companies developing learning-focused AI agents
• Join communities discussing AI in learning and development
• Experiment with basic automation workflows
• Document use cases specific to your organization

Prepare Your Organization

• Identify processes that could benefit from automation
• Collect data about current workflows and pain points
• Build support for AI adoption among stakeholders
• Create guidelines for AI tool evaluation and implementation

Plan for Integration

• Map out potential integration points in your learning ecosystem
• Consider data security and privacy requirements
• Develop success metrics for AI implementation
• Plan for change management and user adoption
• Prepare a governance plan for your educational AI applications

Lead the Change

• Position yourself for success in the AI-enhanced learning landscape:
• Establish yourself as an AI-informed learning strategist
• Build coalitions with IT, HR, and business leaders
• Create pilot programs that demonstrate value
• Develop guidelines for ethical AI use in learning
• Share successes and lessons learned with your professional community

Looking Ahead

The race to make AI agents reliable and cost-effective has already begun. Just as Moore’s Law predicted the exponential growth of computing power, we’re seeing similar scaling in AI capabilities. The cost of AI “intelligence” is decreasing while capabilities are increasing, suggesting that AI agents might become practical “prime time” tools sooner than many of us expect. And this means that those of us who are prepared to leverage these tools will have a significant advantage over our less-prepared competitors.

What Can We Do Now?

• Stay informed and prepared for the next wave of AI innovation
• Start by identifying areas where AI agents could add value to your work
• Experiment with available tools
• Build your understanding of the technology by explaining it to others

Remember, the goal isn’t to replace human learning professionals but to augment our capabilities and free us to focus on the most impactful aspects of our work.

I believe that the future of learning will be a partnership between human expertise and AI capabilities. By starting to explore and experiment now, we can help shape how this technology will best serve our learners, our organizations, and ourselves.

Or we can wait for others to shape our future – and be left behind by those who do.