- The Loop
- Posts
- 🔆 AI Agents - Agentic Systems 101
🔆 AI Agents - Agentic Systems 101
The basics of autonomous systems, USA's updated approach to AI, and the return of TikTok.
Emma G Moller & Allegra Guinan
January 23, 2025
Was this email forwarded to you? Sign up here
🗞️ Issue 54 // ⏱️ Read Time: 9 min
Hello 👋
2025 has barely started, and it's already considered the year of AI agents by industry leaders. While modern agents build on decades of research in autonomous systems, recent advances in language models have transformed them from academic concepts into practical business tools that are reshaping how organizations operate. In part 1 of this series, we introduced Large Action Models (LAMs). Today, we’ll get into the basics of AI agents to help you confidently navigate the AI landscape this year. Welcome to Agentic Systems 101!
In this week's newsletter
What we’re talking about: The evolution of AI agents from their academic foundations to today's practical implementations.
How it’s relevant: 51% of organizations are already exploring AI agents, but most are still grappling with how to implement them responsibly and effectively. As organizations increasingly move from simple chatbots to more autonomous AI systems, understanding both the technical and ethical dimensions of agents becomes crucial
Why it matters: AI agents represent a shift from passive tools to autonomous systems that can reason and act independently, raising critical questions about control, trust, and human oversight. Their development demands careful consideration of ethics, transparency, and the appropriate balance between automation and human agency.
Big tech news of the week…
🇺🇸 President Donald Trump has revoked former President Joe Biden's 2023 executive order on artificial intelligence, fulfilling a campaign promise to repeal regulations that Republicans viewed as hindering AI innovation and imposing "radical leftwing ideas" on the technology's development. The initial purpose behind the regulation was to ensure the safe, secure, and trustworthy development and use of artificial intelligence.
💵 President Trump announced a $500 billion Stargate AI infrastructure project with OpenAI, Oracle, and SoftBank, aiming to establish U.S. leadership in AI, with initial data center construction beginning in Texas.
📱 TikTok has begun restoring service to users in the United States following a brief shutdown on January 18, 2025, in response to a federal law requiring ByteDance to sell its U.S. operations.
🧱 Perplexity launched an API service called Sonar, allowing enterprises and developers to build the startup’s generative AI search tools into their own applications. This launch represents a significant step for Perplexity in the competitive AI search market.
🇨🇳 Chinese AI lab DeepSeek has released an open version of DeepSeek-R1, its so-called reasoning model, that it claims performs as well as OpenAI’s o1 on certain AI benchmarks.
What is an AI Agent?
At their core, AI agents are programs designed to perform tasks autonomously, often with little to no supervision. The smarter the agent, the more complex the tasks it can handle, like booking your next meeting, boosting sales, or identifying market opportunities. What sets AI agents apart from traditional systems is their ability to perceive their environment, make decisions, and take actions toward specific goals. Rather than simply reacting to inputs, agents can plan ahead, learn from experience, and work independently toward objectives - almost like a digital employee.
To understand AI agents, it helps to revisit early distinctions in agency. Weak agency—defined by autonomy, reactivity, proactiveness, and social ability—underpins most modern AI systems. For example, virtual assistants like Siri can respond to your requests but don’t have independent intentions or goals.
In contrast, strong agency involves human-like qualities such as intention, reasoning, and adaptability, driving research in generative AI and conversational agents. These concepts are essential as AI agents now aim to collaborate seamlessly with humans and tackle tasks requiring adaptability and contextual understanding.
Transforming Business Today
Organizations are already seeing impact from AI agents. According to a recent report from LangChain - one of the leading platforms for building AI agent applications - the top use cases are:
Research and Summarization (58% of implementations)
Gathering and synthesizing information from multiple sources
Conducting market and competitive analysis
Summarizing large documents or data sets
Personal Productivity Assistance (53.5%)
Managing schedules and workflows
Processing emails and communications
Automating routine tasks
Customer Service Automation (45.8%)
Handling initial customer inquiries
Routing and escalating complex issues
Providing 24/7 support coverage
Industry analysts predict that in 2025, 25% of companies using generative AI will launch agentic AI pilots or proofs of concept, growing to 50% by 2027.