Riding the AI Wave with George Bandarian: The Dawn of Agentic AI - Part 1

Welcome to the start of an exciting new series where we’ll be exploring the fast-evolving world of Agentic AI. Over the next few weeks, we’ll dive into the current landscape, future possibilities, and key trends shaping this transformative field. For tech innovators and investors, understanding Agentic AI is critical as it promises to reshape industries and unlock new growth opportunities.

What is Agentic AI?

Also known as autonomous AI, Agentic AI represents a major leap forward in artificial intelligence. Unlike traditional AI systems, which rely on specific inputs to function, Agentic AI aims to develop AI agents that can make independent decisions and execute tasks without constant human oversight. In essence, these agents are capable of taking initiative, adapting to new challenges, and working toward goals with minimal intervention.

The Building Blocks of Agentic AI

To fully grasp Agentic AI, let’s break down its core elements:

• Foundation Models: These are Large Language Models (LLMs), which serve as the "brain" of the AI agent. Trained on vast datasets, they enable AI to understand complex topics, generate human-like text, and solve problems in ways that resemble how human experts accumulate knowledge and experience.
• Tool Use: Agentic AI goes beyond giving advice—it can perform tasks. Imagine an AI assistant that can search the web, send emails, or edit documents all on its own, taking action based on the context of the task at hand.
• Memory: Just as people learn from past experiences, Agentic AI systems can recall previous interactions, remember user preferences, and apply past learnings to new tasks. This allows them to become more personalized and efficient over time.
• Multimodal Capabilities: While early AI mainly dealt with text, Agentic AI can interpret images, understand speech, and even generate visual or audio content. This expanded capacity brings the AI closer to having human-like senses and abilities to create.

The Agentic AI Ecosystem

The rise of Agentic AI has given birth to a variety of applications:

• Horizontal Applications: AI agents are used across industries for tasks like customer service or automating sales processes.
• Vertical Applications: These are industry-specific uses of AI agents, such as in healthcare, finance, or manufacturing.
• Agent Infrastructure: The platforms and tools that help create and deploy AI agents are crucial to the growth of this space.

Two Strategic Approaches to Agentic AI

As companies race to build the most capable AI agents, they typically take one of two approaches:

1. Model-First Startups

These startups focus on creating advanced AI models, believing the foundation model itself will determine the agent’s success. They are pouring resources into developing more sophisticated LLMs to power Agentic AI.

Key players include:

• OpenAI ($13B raised): Known for GPT models and working on AI agents that control user computers.
• Anthropic ($7.3B raised): Similar strategy, with active research into AI agents.
• Adept ($413M raised): Specializing in models trained on user actions.
• Imbue and Magic AI: Focusing on AI agents for coding and software engineering.

The big question these companies face: which model will prove most effective for building true AI agents?

2. Workflow-Focused Startups

These companies leverage existing AI models and concentrate on applying them to specific tasks or industries. They believe success comes from optimizing how AI is applied, not necessarily creating a new model from scratch.

Examples include:

• Harvey: Developing AI agents for the legal sector.
• Cognition Labs: Specializing in automating coding tasks.
• Lindy: Offering platforms that handle a variety of tasks like scheduling and note-taking.

The Promise of Agentic AI

While the field is still young, the potential of Agentic AI is enormous:

• Boosting Productivity: AI agents can handle complex, multi-step tasks, freeing humans to focus on higher-level strategic work.
• Personalized Assistance: Agentic AI can learn and adapt to individual needs, offering highly customized support.
• Driving Innovation: By autonomously exploring problems and running experiments, these agents can spark breakthroughs across industries.
• Enhanced Decision-Making: By analyzing vast amounts of data, Agentic AI can help businesses make smarter, data-driven decisions.

What’s Next?

As we wrap up this introduction to Agentic AI, it’s clear we’re on the brink of a major shift. In upcoming articles, we’ll take a deeper look into how these AI agents are already transforming industries, the technologies driving this shift, and how startups and investors can position themselves to capitalize on this trend.

At Untapped Ventures, we’re excited about the vast potential of Agentic AI to create new markets and revolutionize existing ones. We’re actively investing in innovative startups that are pushing the boundaries in this space, as we believe they will be key to shaping the future of AI and its impact on society.

Stay tuned for the next article where we’ll explore real-world use cases of Agentic AI and the companies leading the charge.

As AI pioneer Andrew Ng put it: “AI agent workflows will drive massive progress this year, perhaps more than the next generation of foundation models. This trend is critical for anyone in AI to watch.”

If you're building in the Agentic AI space, don’t miss the opportunity—apply now to pitch your startup for funding and join us in shaping the future!