One of my favorite storybook series that I have read for my son when he was young was this Alphapets series. Beyond the creative naming of the animals, I was fascinated by the characteristic or adjective associated with each one. Rupert the Resourceful Rhino, Fenton the Fearful Frog, or Delilah the Demanding Duck! Even today, we jokingly revisit these names to see how many we can recollect—the name, the trait, and the animal.
You might wonder how this relates to AI Agents. Don’t worry—we’ll explore that connection shortly.
What are AI Agents?
This topic has recently generated significant buzz. AI agents have become the latest trend, thanks to ChatGPT and its counterparts like Gemini, Claude, and Llama.
“I think we are living in a world where there are going to be hundreds of millions of AI agents … maybe there are more AI agents than people in the world.” – Mark Zuckerberg
“Agents are the new apps, and someday, there will be thousands of them.” – Forbes
But what exactly are they?
To answer this, let’s briefly step back and answer this question.
What is AI?
Intelligence, whether human or artificial, is the ability to learn from training or experience and apply that knowledge to adapt to situations, solve problems, or provide answers. Simply put, it’s the capacity to learn, understand, think, and act and the ability to be useful to oneself and to others.
Intelligence is only valuable when put it to use. It’s raw power that is waiting to be harnessed. This principle applies to the artificial version (AI) as well, which is why AI Agents are designed as software incarnations capable of understanding human requests, follow pre-defined routines – predominantly codified – and perform autonomous tasks that generate value.
This value manifests itself as enhanced user experience, quicker responses, time efficiency, and improved productivity and thereby in value generation. Moreover, technology enables the delivery of this intelligence at a scale that humans would find quite challenging.
Hang on. Haven’t we heard this before?
Weren’t chatbots or the RPA (Robotic Process Automation) solutions, once the darlings of the tech world, supposed to do the same thing? Are AI Agents just chatbots reborn with a fancy new name?
Good question.
Indeed, AI Agents share many similarities with their predecessors, chatbots and RPA solutions, in both appearance and function. At their core, AI Agents feature a clear, primarily conversational interface. This could take the form of a chat interface like the popular ChatGPT, a voice assistant, a telephone call, messaging platforms or integration with collaboration platforms such as Slack or Teams.
I would even say that many chatbots excelled only in their interface. Most relied on pre-defined dialogue patterns and decision trees, resulting in weak language understanding (LU) capabilities and a subpar user experience.
And, Robotic Process Automation (RPA) solutions helped automate repetitive human tasks. You define the tasks and the rules and the automations perform those with 100% reliability and speed.
With an AI Agent, three key aspects operate behind the scenes. Remember, the AI Agent is still a piece of software:
- Understand: As a primarily conversational interface, one crucial feature is excellent Language Understanding (LU) capability. This allows the agent to clearly comprehend what’s being asked.
- Think: This constitutes the core AI component and determines the course of action.
- Act: This is the “doing” part. AI Agents can perform a wide variety of actions, such as answering queries, processing business transactions, or escalating complex issues to human agents. The aim is to provide a seamless experience, ensuring customers receive timely and useful responses.
With AI Agents, there is a significant jump in the value creation in terms of better language understanding capabilities, to take complex actions and to become smarter with continued usage.
How is that possible?
Understand
The use of advanced language processing techniques offered by Large Language Models (LLMs) has been a significant step forward in understanding and responding to user inputs. This has vastly improved Language Understanding (LU) capabilities and the ability to understand the context and generate responses in more human-like language constructs. LLMs or even the business domain focused Small Language Models (SLMs) offer a more practical approach for enterprises and industry focused solutions.
Think
This is the core brain behind an agent.
The AI Agent now has the ability to process input questions across large volumes of data and suggest answers and solutions that would make sense. Additionally, they incorporate “guardrails”—a term borrowed from physical safety barriers that prevent falling off cliffs.
In AI, guardrails are a set of constraints that keep the AI operating within certain boundaries—technically, legally, and ethically. They prevent the AI from causing harm or making incorrect predictions or outcomes. These guardrails are defined using prompts or data.
If you want to contrast an AI Agent with a chatbot, it’s this “thinking” step that makes the biggest difference. Any agent without the ability to think or use AI is merely a simple workflow tool.
Remember, an AI Agent is designed to improve over time through continued usage and develop a “memory”.
Act
This is the most crucial step for an Agent, whether Human or AI—to deliver meaningful outcomes. These could include answering questions, summarizing documents, or booking appointments or quite simply, integrating many of the erstwhile RPA automation . In addition, it might also involve escalating complex queries to a human when the AI Agent can’t provide an answer.
With the wide array of webhooks, REST APIs and business automation software like Zapier available for various business applications and functions, the possibilities for what an AI Agent can do are virtually limitless.
So, that’s your AI Agent demystified at the 10K foot level.
Use Cases
In terms of the use cases, over the past many years, plenty has been written already on the application of chatbots, voice assistants and RPA solutions.
With respect to this new incarnation, AI Agents, two primary factors need to be considered while picking the right use case or the right business problem to be solved.
First, focus on small, targeted applications for specific industries and domains.
While Large Language Models (LLMs) like GPT and BERT have seen their parameter counts skyrocket into the hundreds of billions, this trend has also sparked interest in Small Language Models (SLMs). These SLMs are custom-built for specific business and industry domains, offering greater precision and cost-efficiency in terms of computing power. Their applications span various sectors, including IT, legal, real estate, and healthcare.
Second, focus internally. A common misconception about the usage of Chatbots and the Conversational AI technology, is that they’re primarily used for external or customer-facing use cases.
Remember, AI’s primary role is to enhance or augment human capabilities and help us perform our jobs more effectively to be more efficient and at scale.
While the erstwhile chatbots do offer 24×7 availability, at scale to external users, I believe some of the most valuable applications of AI Agents could be for internal users, staff, and human agents.
Harness the power of a well-crafted AI agent to serve as an internal mentor, act as a knowledgeable coach, dissect large documents, streamline tasks, schedule appointments, or provide highly context-specific answers on demand.
Imagine a diverse array of AI Agents skilled in various tasks working closely with everyone across the enterprise. You can simply hire an agent or craft your own custom assistant.
As Zuckerberg suggests, while we may not see millions of AI Agents, we could certainly benefit from a significant number of purposeful ones designed to assist us with various tasks.
The potential outcomes are plenty: saving time, reducing error rates, making knowledge accessible at scale, decreasing wait times, and eliminating unnecessary call transfers.
Summary
Let’s go back to the topic of Alphapets.
How easy would it be for someone to create similar AI Agents, with a character? It is not about the dream of bringing in anthropomorphic characters to the software applications. The Alphapets metaphor kinda offers a better perspective on the value add from AI for humans.
Erika the Empathetic HR Specialist – that can provide instant answers on all the HR policies, provide a self-service interface for mundane tasks and work in tandem with a human HR specialist for complex questions.
While Erika could serve as the first level of support for your staff, consider a more innovative approach. Imagine a new specialist joining your HR team. Erika can assist this newcomer in providing exceptional service to your staff by acting as a coach on company policies. In this way, Erika truly augments human capability, enhancing the skills of your HR professionals.
Eva the Enterprising Auto Dealer – who knows the best about your automobile, schedule service appointments with proactive reminders and alerts. With Eva connected to your Dealer Management System (DMS), it can support your service delivery team by proactively notifying them of upcoming service schedules and aiding them in their customer success journey.
Or, Lego, the Resourceful Reference—an AI agent capable of analyzing extensive legal documents, providing summaries of various terms and conditions, and sharing that information with others. Need to review a new legal document or have a question about a specific clause? Lego can help by summarizing documents, answering questions, and analyzing content, thereby saving you a significant amount of productive time.
That’s the world we envision on the AI horizon: democratization of AI powered automation and the ability to create AI-powered assistants for a diverse set of users and use cases without writing code and that doesn’t take months to implement.
A bunch of AI Assistants that you can build or hire to work in tandem with humans, to augment human capabilities to match the agility and need of the new world we are in. These assistants will help us perform our jobs more effectively, offer higher-quality service, and—most importantly—delight both our customers and ourselves. And, a platform ecosystem that helps solve the last mile problem with AI for meaningful outcome.
ParrotGPT makes that possible.
How can we build an AI Agent for you?
