The Path of an AI Creator News: AI agents will change the customer experience! What's the secret to improving customer satisfaction? #AIAgent #CX #CustomerExperience
Video explanation
That's amazing! Will "Agent AI" change the future with customers?
Hello, I'm John, your guide to the world of AI and technology! Recently, you've probably seen the word "AI" on TV and the Internet a lot. Some of you may have had a conversation with a smart AI, like "Chat GPT."
Today, we will introduce a type of AI that is attracting particular attention.Agent AI" is the technology that we use to provide the service we experience as customers.Customer Experience" or "CX"). As always, I'll explain it in an easy-to-understand way!
What is "Agent AI"? How is it different from regular AI?
First of all, what is "agent AI"?
The AI you often hear about, such as AI that creates sentences or answers questions, is often called an "LLM (large-scale language model)." This is like a "language doctor" that gets smarter by reading a lot of sentences.
"Agent AI" can take this "word doctor" ability a step further.Think, judge and carry out tasks independentlyIt's like having an excellent secretary for each of us. When we give instructions, it gathers the necessary information, connects with other systems, and does various other tasks to achieve our goals.
In fact, agent AI is already being used in companies to help employees with their work: for example, helping human resources departments find new employees, helping marketing departments think up the perfect advertisements for each customer, and helping IT departments answer questions about computer problems.
From now on, customer experiences will evolve with "Agent AI"!
This trend is likely to spread not only to employees, but also to us customers. Experts say, "It's only a matter of time before agent AI becomes the new norm for customer experiences."
What this means is that you may find yourself less likely to get frustrated with confusing operations on store websites, struggle to find what you're looking for among a sea of information, or get tired of long input forms, as you have in the past.
Instead, the agent AI will quickly understand our preferences and the situation, making interactions easier and smoother.
John Kim, CEO of a company called Sendbird, said, "Companies will introduce AI agents with expertise in various areas, such as product knowledge, inventory, pricing, shipping, and legal issues. This is already happening in the retail industry, where AI makes personalized recommendations and proactive support to make shopping more enjoyable. In the future, each of us may have AI assistants that specialize in various areas, such as money, hobbies, health, travel, etc."
The first step is to leave the "boring work" to us!
That being said, it can be a bit worrying to suddenly entrust everything to AI. In fact, there have been cases in the past where companies rushed into introducing AI, which ended up affecting their customers and the company's reputation.
That's why many companies are taking a cautious approach to using AI agents in their customer experience. Specifically, they are creating rules for using AI correctly (known as "AI Governance" (He says) it's important to take preparations such as improving the quality of the data the AI uses to learn and conducting thorough testing.
So where do you start? Experts say:Why not start with the repetitive tasks that are frustrating for your customers and boring for your company?” he advises.
Dave Singer of Verint says, "Generative and agent AI are most effective at automating the most tedious, repetitive, and detailed tasks in customer interactions." For example,
- Ask the right questions at the beginning of each conversation to understand the situation
- Find answers to your questions
- Handle paperwork after the phone call
By offloading these tasks to specialized AI bots, human staff can focus on more important tasks, improving customer experience, reducing costs, and increasing sales.
Other examples include manuals and help pages that you look at when learning how to use a product, installing it, or resolving problems. Instead of reading a thick manual from cover to cover, you might be able to find the answer faster and easier by asking an AI agent.
John Kennedy, CTO of Quickbase, suggests, "Let's think about how we can use generative AI and LLM to improve the places that customers often visit when using the product, such as product help pages, user-created information sites (Wikis), and online communities." It would also be a good idea to have AI quickly produce templates tailored to customers by industry or role, or to have AI guide them on their next steps.
Dion Nicholas, founder of a company called Forethought, is not just interested in finding information quickly, but also in having AI perform simple tasks. "One of the easiest things to develop using LLM is a chatbot that can quickly find information from FAQs (frequently asked questions). But if you incorporate an agent AI into a website or app, you can expect even greater benefits, as it can take actions such as resetting passwords on behalf of customers or checking the status of their orders," he says.
The key to utilizing AI is "centralizing customer data"
To realize a more interactive customer experience with AI, we need to make sure that the AI is learning and working correctly.Customer data is neatly organized and collected in one placeTo that end, companies are using technologies such as "Customer Data Platforms (CDPs)" and "Data Fabrics" to connect customer data and past interactions.
Tara Desao of Pega says, "A powerful customer experience strategy using AI is meaningless if the underlying data and its management system are not solid. To keep the data fresh and accurate, you should emphasize continuous testing and learning strategies." This not only improves the performance of AI, but also reduces risks and increases customer trust.
When centralizing customer data, it is necessary to thoroughly take care of security, authority management (i.e. who can access what information), identity verification, and so on.
Osmar Olivo of a company called Inrupt says, "By rethinking how data is stored and accessed, and transforming information that was previously managed in a fragmented way into a user-centric data model, we can create smoother, more responsive web and mobile experiences that adapt to individual preferences in real time." It's important that the information and suggestions provided by AI are trained using a variety of actual data, while allowing users to make corrections themselves, saying "this isn't right" or "this is better," or to communicate their own preferences.
However, according to Manish Rai of a company called SnapLogic, more than 8% of generative AI projects fail due to issues with data connectivity, quality, and reliability: "Success will depend on tools that make it easy to develop agents, make the data AI-ready, and ensure reliability by monitoring and evaluating accuracy and compliance with rules."
Rosalia Siripo of the company KNIME said that in many cases, there is a process where a human checks whether the results produced by the AI agent are correct ("Human in the Loop"In some cases, we have a dedicated AI agent that will check the results and, if we're not happy with the results, we'll ask you to do it again."
Could AI agents help with frustrating phone calls and chats?
Not only does it automate information searches and simple tasks, but it also makes phone and chat inquiries stressful for both customers and human operators. In one survey, 23% of people said they would rather watch paint dry than experience bad customer service over and over again. That's tough!
Up until now, there have been many chatbots that could only answer questions according to set rules, but in the future, AI agents that can respond flexibly based on data will appear. And with the support of AI agents, human operators will be able to focus on solving more complex and difficult problems.
"There's a clear link between customer satisfaction and effective self-service usage," says Cisco's Vinod Muthukrishnan. "The evolution to true agent AI will transform the self-service experience by orchestrating the entire interaction between your brand and your customers. These advanced AI capabilities will empower customer-facing teams to deliver smarter, smoother interactions that meet customers where they are and at their convenience."
The challenge here is the data issue, and the other is that customer experiences to date have often been created as "point" solutions that only address a part of the overall customer behavior. Experts say that when changing to experiences that use generative AI,Design thinking"They recommend that companies adopt a more holistic, connected approach to redesigning their experiences to be more holistic and connected," he says.
Chris Arnold of the company ASAPP explains, "Websites, mobile apps, and business-to-consumer messaging apps are typically connected behind the scenes to customer-specific data sources that can answer questions and solve problems. Leveraging LLM to deliver a conversational, personalized experience goes far beyond the transactional experience these apps provide on their own."
Before introducing an AI agent, thorough testing is essential!
For companies looking to develop more advanced customer experiences or autonomous AI agents, this is a great way to validate their capabilities.Comprehensive Test Planis absolutely essential.
Mechanisms to avoid inappropriate or off-topic conversations (e.g., prompt filters, AI response moderation, content safeguards, etc.) are important, but beyond that, companies need to ensure that their AI agents respond appropriately, accurately, and ethically.
Miles Ward, CTO of a company called Sada, strongly emphasizes that "we wouldn't ship an agent to the public without testing and monitoring it. Rigorous testing for accuracy and performance is a non-negotiable. If we aren't confident that it provides a smooth and reliable experience, we're just creating new problems."
Ganesh Sankaralingam of the company LatentView says that AI experiences and LLM responses should be tested for accuracy and performance across five dimensions:
- Relevance: How relevant the response is to the question.
- Groundedness: Is the response based on the information or data entered?
- Similarity: How close the AI-generated response is to the expected model answer.
- Coherence: Does the response flow naturally and use human-like language?
- Fluency: Is the response grammatically correct and uses appropriate vocabulary?
Dion Nicholas of Forethought adds: "Businesses should test their AI agents on questions from past customers to see how they respond. They should also measure the percentage of customer interactions the AI can handle autonomously, and use other evaluation models to check the tone and accuracy of the conversations."
What is the future of customer experience that Agent AI will pave the way for?
So how will agent AI change our customer experiences in the near future? Mo Cherif of Sitecore suggests that "to create a truly agent-like experience, we should not just make the existing one a little better, but build an experience based on generative AI from the beginning."
There are a few different views on how agent AI will evolve in the future.
Some foresee a future where AI agents will become more autonomous and humans will trust them to make more complex decisions and take a wider range of actions, while others foresee a more human-centric approach, where AI agents will augment human capabilities and act as partners to make smarter, faster and safer decisions.
Michael Wallace of Amazon Web Services (AWS) talks about the possibility of agent AI being able to solve problems without human intervention. For example, in the future, contact centers will be able to detect problems on their own when something goes wrong, automatically reallocate resources, update communications with customers, and resolve issues before customers even realize they have a problem.
"Imagine this: an airline faces a sudden increase in traffic due to bad weather," says Wallace. "With Agent AI, the contact centre could make autonomous decisions about rebooking or proactively notifying passengers, freeing up human agents to focus on addressing complex customer needs rather than administrative tasks."
"AI shouldn't just be used to automate customer experiences. It should make those experiences more human and more intelligent," says Doug Gilbert of Sutherland Global. "The real value of generative AI is not in replacing human interactions, but in augmenting them to make them smarter, faster and more natural. The secret is AI that learns from real-world interactions and is constantly evolving to feel less robotic and more intuitive."
It is likely that we will see the realization of both fully autonomous AI agents and AI agents with human support. Until then, it is important for companies to thoroughly research customer needs, improve the quality of data, and establish rigorous testing systems.
A word from John
Wow, it seems like an exciting future awaits! Perhaps the day will come when we will be freed from the hassle of complicated procedures and the frustration of never getting through to customer service desks. Of course, AI won't do everything for us, and it will be important for us humans to know how to interact with it well. But I have high hopes for the potential for AI to make our lives more convenient and comfortable!
This article is based on the following original articles and is summarized from the author's perspective:
Design a better customer experience with agentic AI