To Agent or Not to Agent?

Introduction to AI Agents

Adel Nehme

VP of AI Curriculum, DataCamp

A Tale of Two Customer Support Teams

customer_support_a

  • "How do I track my order?"
  • "How do I return an item?"
  • "How do I change my shipping address?"

customer_support_b

  • "I was charged twice, but one order was cancelled, and I have store credit from a previous return that wasn't applied correctly"
1 Image generated with GPT-4o
Introduction to AI Agents

The Tale of Customer Support Team A

  • "How do I track my order?"
  • "How do I return an item?"
  • "How do I change my shipping address?"

 

Problem qualities

  • They require simple decision-making
  • They do not require accessing customer information and history
  • They all have discrete, predictable answers
Introduction to AI Agents

The Tale of Customer Support Team B

"I was charged twice, but one order was cancelled, and I have store credit from a previous return that wasn't applied correctly."

Problem qualities

  • They require complex decision-making
  • They require accessing customer information and history
  • They require adaptive solutions
Introduction to AI Agents

A Tale of Two Customer Support Teams

customer_support_a

  • A simple chatbot trained on pre-trained knowledge.
  • No tool use or actions required.

customer_support_b

  • Agentic solution with tool access
  • Can generate remediation strategies and update systems
1 Images generated with GPT-4o
Introduction to AI Agents

When to Use AI Agents

Criteria for using AI Agents

  1. Require complex decision-making
  2. Require heavy reliance on unstructured data
  3. Have difficult to maintain rules
  4. Require adaptive problem solving

Examples of agentic use cases

  1. Autonomous customer support systems
  2. Coding assistants that can read code bases, provide updates, and implement them automatically
  3. A deep research assistant that can synthesize research
Introduction to AI Agents

The AI Agents Tooling Ecosystem

Introduction to AI Agents

The AI Agents Tooling Ecosystem

$$

AI_agents_1

Introduction to AI Agents

The AI Agents Tooling Ecosystem

$$

AI_agents_2

Introduction to AI Agents

The AI Agents Tooling Ecosystem

$$

AI_agents_3

Introduction to AI Agents

The AI Agents Tooling Ecosystem

$$

AI_agents_4

Introduction to AI Agents

Build vs Buy: A Framework

Off-the-Shelf Tools
Tackling a specific domain or use-case (e.g. AI-assisted coding)
Mature, well-tested solution already exists in market
Want to minimize maintenance overhead
Introduction to AI Agents

Build vs Buy: A Framework

Off-the-Shelf Tools Low-Code/No-Code Platforms
Tackling a specific domain or use-case (e.g. AI-assisted coding) Need some customization but not complete control
Mature, well-tested solution already exists in market Workflows are moderately complex but follow common patterns
Want to minimize maintenance overhead Want business users to modify the agent without engineering help
Need to integrate with existing systems quickly
Introduction to AI Agents

Build vs Buy: A Framework

Off-the-Shelf Tools Low-Code/No-Code Platforms Agent Frameworks (Build)
Tackling a specific domain or use-case (e.g. AI-assisted coding) Need some customization but not complete control Use case involves proprietary systems
Mature, well-tested solution already exists in market Workflows are moderately complex but follow common patterns Handling sensitive data
Want to minimize maintenance overhead Want business users to modify the agent without engineering help Agent is core to competitive advantage
Need to integrate with existing systems quickly No existing solution meets specialized requirements
Need complete control over agent's behavior and evolution
Introduction to AI Agents

Let's Practice!

Introduction to AI Agents

Preparing Video For Download...