What is Agentic AI? The new keyword to watch out for in AI
18 Oct 2024
3.5 min read
Agentic AI is the latest buzzword doing the rounds in the world of Artificial Intelligence. This is a guide to what it means, and what the implications are for customer service leaders.
Agentic AI refers to an artificial intelligence system that can act on its own and control actions. The idea is that it is self-driven, and able to make decisions and learn from previous decisions and behavior to change and optimize. Agentic AI systems are often powered by large language models and other advanced ai models, which underpin their autonomy and decision-making capabilities. These systems are designed to operate autonomously in complex environments and tackle complex problems that require sophisticated reasoning. Put simply it means being able to act like an agent, to make decisions and take actions in order to achieve a goal.
If this is all new to you, it probably sounds futuristic, but in reality this is available today, to a large extent. We’ll use examples from customer service AI to illustrate how it works.
AI Agents vs. AI Assistants
Sometimes you will hear AI Agents spoken of interchangeably with AI Assistants, partly because Agents and Assistants are both people who work for you and help you. But they are not quite the same.
AI Assistants are tools that you can query with a specific request, such as “Summarise this conversation”, or “Tell me what the differences are between these products”. The key is that it is responding to a specific prompt or specific task.
AI Agents, on the other hand, have more…well, agency. This means they are more proactive, can act independently with minimal human intervention, and can look to achieve a certain goal or set of outcomes.
For example, if you asked both an AI Assistant and an AI Agent to buy a gift for your sister, the AI Assistant would respond to the specific task you provide, such as searching for a particular item you specify. In contrast, the AI Agent would be able to plan and execute the process autonomously—finding something it thought your sister would like and ensuring it got there on time.
How agentic AI would deal with customer service requests
Imagine that a customer asked “What materials are your trainers made from?”. In the past, if a chatbot was asked that it would have required training to recognise that specific question, and then a templated response would be given. This could be one of several templated responses. The point being that the AI would ‘understand’ the question and then trigger a programmed response.
Currently, with generative AI, it can be a bit more sophisticated. The AI still has to recognise the question, but it can then be trained to look for the answer to the question in the knowledge base. This means that the answer can always be “live” - i.e. if the knowledge base is updated, the chatbot does not need to be re-programmed. But crucially the AI needs to have been trained to only look in certain places to find the answer, so it doesn’t hallucinate.
So far, so good, but these are only helping with relatively simple question and answer responses.
What about if a customer asks the chatbot to do something, such as amend an order? This is possible currently if you have the deep integrations to the order management system required. Once the request is understood, then the AI can look up the order, and then overwrite it with the new changes. Agentic AI can perform tasks and execute tasks by directly interacting with third party applications and external tools through application programming interfaces, enabling real-time actions and automation. Additionally, agentic AI can process data from various sources and automate repetitive tasks in customer service workflows, improving efficiency and reducing manual workload.
This is something that DigitalGenius customers such as Organic Basics will be familiar with.
Agentic AI becomes self-improving
The example outlined still requires some flow mapping in most cases. Creating an if-then series of steps allows the retailer to put the AI down strictly defined paths.
The next step of AI is to stop strictly defining paths and allowing the AI to be guided by what it thinks is the best step.
The idea of an autonomous AI running wild with little oversight, making its own judgements and deciding things for itself is a little bit terrifying. It’s HAL from 2001: A Space Odyssey-level stuff.
But just as with current generative AI systems there would need to be guardrails, and lines that an agentic AI would not cross. Human oversight remains crucial to ensure safety, ethical compliance, and accountability, as a lack of constant human oversight in autonomous agentic systems can introduce risks if the AI manages complex problem solving tasks independently. Plus it would begin to have a wealth of experience to draw upon regarding the best course of action at each stage.
Agentic AI systems improve their performance over time by using reinforcement learning and other techniques to enhance their problem solving abilities. Through these processes, AI agents learn from experience, optimize their actions, and refine their capabilities as they interact with data and environments.
For example, the thing that would make customers happiest is probably that they always get a refund, even if there hasn’t been a problem. An AI that issues refunds to all customers would be a disaster – so this could be somewhere where strict rules are put in place. For example, only issue a refund to a loyal customer who has never had an issue before, otherwise pass to a human.
Adding proactivity
One of the key differences between AI Assistants and AI Agents is proactivity.
The idea is that the AI is constantly looking for issues that need resolving, such as packages that are late. Agentic AI achieves this by collecting data from multiple data sources, analyzing data in real time, and identifying relevant information to anticipate problems. In the case of Porto’s for example, the AI Agent is searching to see whether the packages, filled with perishable bakery goods are going to be delivered on time, because otherwise the packages will need to be replaced.
An Assistant would instead be waiting for the customer to realise that the package was delayed and to ask for a replacement or a refund. This is yet another way that Agentic AI, or AI Agents are being deployed, as this proactive approach enables agentic AI to manage complex workflows in customer service.
Getting started with agentic AI
The ability for AI to be truly autonomous is on its way. But what we are likely to see are vertical AI Agents – that is AI agents that are proficient in particular areas, such as ecommerce.
Implementing agentic AI in business contexts involves deploying agentic AI tools and AI powered agents to automate enterprise software and manage complex tasks and complex processes that traditionally require significant human input. Agentic AI systems can operate with a higher degree of autonomy, directly interact with diverse sources such as APIs, databases, and cloud services, and are already being used in real world applications across industries. For example, a few examples of multi agent systems and agentic AI systems include automating decision making in supply chain management, orchestrating enterprise software for fraud detection, and tackling complex challenges in healthcare by analyzing patient data.
That’s exactly what DigitalGenius is. Our platform is built with deep integrations into your tech stack, and when trained on your business processes can act like an AI Assistant, and move towards an AI Agent, proactively checking for issues and resolving them before customers realise.
If you want to get a sense of the future of AI in ecommerce, talk to DigitalGenius today.
