agents... agents everywhere

Exploring the hype and recent developments in AI agents - MLforSEO Newsletter #001

Hi 👋🏻

2025 has kickstarted the era of AI agents.

Well, not really - the concept of an AI agent has been around since the 50s but the latest surge in generative AI, deep learning, and orchestration are what propels this concept forward.

In this newsletter edition, let's explore AI agents together, and try to imagine what this automation would lead to for SEO and digital marketing?

what are agents?

Agents combine reasoning, logic, and access to external information that are all connected to a Generative AI model. Think of them like ChatGPT but better - personalised assistant with access real-time information, that can suggest real-world actions, and plan and execute complex tasks autonomously.

Agents are as powerful as their tools (functions), extensions (API access), and data access, and can interact with both structured and unstructured data.

When implementing agents, the folk at Anthropic recommend that you follow these three core principles:

Maintain simplicity in your agent's design.
Prioritize transparency by explicitly showing the agent’s planning steps.
Carefully craft your agent-computer interface (ACI) through thorough tool documentation and testing.

Agents Whitepaper by Google | Kaggle

Building effective agents \ Anthropic

Navigating the AI Frontier: A Primer on the Evolution and Impact of AI Agents - World Economic Forum

There are different types of agents, with general purpose and Retrieval-Augmented Generation (RAG) gaining prominence quickly.

Some use-cases for SEOs and marketers:

Hyper-personalised content - By accessing and analysing consumer behaviour data, RAG agents can generate content tailored to individual preferences, significantly boosting engagement and conversions.
Example: An e-commerce brand uses agents to write hyper-personalised emails personalised for different customer segments.
Audience understanding - By gaining access to databases and analysing data on query semantics, customer acquisition data, and user behaviour data from tools like Microsoft Clarity, a RAG agent can provide granular insights into high-value search terms and user journeys.
Example: A SaaS company uses agents to identify long-tail queries like best project management tools under $50 paired with Clarity’s session recordings to pinpoint where users drop off, giving actionable and specific recommendations on landing page design for improved conversions and user experience.

Explore more with these resources, and learn the latest news around AI agents:

OpenAI launches tasks scheduling (BETA)

AutoGen v0.4: Reimagining the foundation of agentic AI for scale, extensibility, and robustness - Microsoft Research

Introducing ambient agents - LangChain release

Announcing the General Availability of the Contextual AI Platform - Contextual AI

fab or fad?

AI agents essentially promise to solve the biggest limitations of LLMs - their inability to access data, act autonomously and effectively integrate with external systems to execute complex tasks. Combining reason with knowledge ✨

While many claim AI agents will be the biggest things of 2025, others are doubtful.

AI agents are so far implemented at a pretty small scale (yet showing somewhat of a promise) for tasks like data retrieval, data analysis, and data organisation in areas like customer service or administration.

Others think that AI agents will hit a wall due to factors like costs, computing resources, orchestration complexity, and authorisation. Aspects like data protection, security, privacy, and ethics might also hinder wide-spread adoption and use.

Here are a few videos that introduced me to new perspectives: