Prompted: Why Agents Are the Next Big Thing in AI

Just when we’ve finally wrapped our heads around AI chatbots and large language models, the spotlight has shifted to something even more ambitious: AI agents. Tech giants from OpenAI to Google have recently made major announcements about their advancements in AI agent technology. The buzz around the development of AI agents has been growing, being labeled the “next step” in AI’s evolution. With the potential for AI agents to make a significant impact across various business functions, the term “AI Employees” is quickly gaining traction.

So what sets agents apart from the AI tools we’ve come to know (and sometimes love)? And can they really deliver on their promises to transform the way our marketing teams operate? Let’s dive in.

AI Agents: The Next Evolution

Marketers are now accustomed to using ChatGPT throughout our work days, but AI agents are made to do more than just answer questions or generate code. While chatbots are reactive (waiting for you to ask them something, then responding), AI agents are capable of taking initiative based on the environment, inputs, and predefined goals. And, they can interact with the world around them. For example, picture an AI that could browse through a conference agenda and suggest relevant sessions to attend, create reminders on your calendar, and draft personalized follow ups for connections you made at the event.

The call for AI agents is coming from… everywhere. In early October, OpenAI launched Swarm Framework, an open-source project aimed at helping developers create their own agentic AI systems. While the release was perceived as “low-key” by OpenAI update-watchers, the resulting swarm of announcements came out in high gear. 

Anthropic announced AI agents that can interpret what’s on a computer screen, select buttons, enter text, navigate websites and execute tasks through any software and real-time internet browsing. Jared Kaplan, Anthropic’s chief science officer, said the tech can “use computers in basically the same way that we do.” Microsoft is also actively working on bringing agents into their products, categorizing agents into business functions like sales, operations and service. Google’s “Project Jarvis,” which can do everything from gathering research to booking a flight, is set to be released before the end of the year.

Testing the AI Agents

With the flood of announcements, it looks like AI agents are here in force. But can they really perform human-level tasks by themselves? AI expert Ethan Mollick recently published a blog post on his test of Anthropic's AI agent, Claude. The professor used the technology to play an online game, Paperclick Clicker, prompting only with the URL of the game and the instruction to “win.” This test, said Mollick, was “a good illustration of the strengths and weaknesses of these early agents.” While the agent displayed resilience and strategic thinking throughout the game, including A/B testing, it also strayed towards common AI pitfalls: stubbornness and a susceptibility to spiraling after minor mistakes, demonstrating challenges in speed and efficiency.

Screenshot from Ethan Mollick's test of Claude. The left is the AI agent, and the right is the desktop it was controlling.

Not Quite There Yet

While AI agents offer a lot of promises, the tech isn’t quite there yet. According to Gartner analyst Arun Chandrasekaran, “We’re in the super super early stage of agents.” Chandrasekaran sees agentic AI as “one of the key research goals of AI companies and research labs in the long run.” AI models will have to significantly evolve before they can enact true autonomous decision-making. To perform complicated tasks AI will need reasoning, memory and “the ability to remember and contextualize things,” according to Chandrasekaran.

What’s Next for Businesses

While the tech may still be in development, marketing teams are already starting to integrate simpler, task-based AI agents into their software workstreams. Hubspot introduced content agents that can scale a user’s content creation to amplify their marketing efforts. Salesforce is bringing AI agents into the customer service experience, positioning the success of this technology dependent on the partnership between the AI agent and the human. 

Even if you’re an AI skeptic, this technology is worth keeping an eye on. The vision of AI agents reshaping marketing operations may be closer than you think. Understanding how to work with and integrate these AI agents as technology advances will be essential for maintaining a competitive advantage in today’s business landscape.

Stay up to date with the latest in AI with Prompted, our biweekly blog series bringing you relevant updates and insights on how to strategically add AI-driven solutions into your marketing toolkit.