AI, But Make it Human

By Machielle Thomas
Human agent smiling

So you’ve started using generative AI to help improve your research papers, customer services or organize your daily calendar. Does it sound a little too Jetsons-esque? But here’s the thing: AI agents have the potential to sound as human as you and I. It’s all about understanding the nuances of human communication. People have nuisance, personality, voice, and tone, and there’s nothing worse than seeing a lapse in conversation that feels unnatural.  Most generative AI is built with a large language model (LLM) that is designed to understand and generate human language. However, it’s crucial to note that these models can be trained using large datasets like books, articles, websites, or social media to mimic humans, but they don’t “understand” language in the same way humans can. The urgency lies in developing AI that truly comprehends language, just like we do. 

Think back to when you were a student and crunching a bunch of chapters from your textbook right before a test. You may grasp the general concept, but it’s safe to say you may not fully comprehend the total weight of the content in less than 24 hours. That disconnect between the content you are reading versus understanding can be applied to why our generative AI can present itself as human but doesn’t sound like a human. Turning your generative AI into a human-centered AI may sound like a tall order, but it’s a crucial step in enhancing customer experiences.

A human-centric customer experience isn’t just about being kind and solving problems—it’s about how effortlessly they get solved. One Grammarly user recently shared their experience with an AI agent (powered by Forethought) that turned what could have been a frustrating refund request into a seamless interaction. In under two minutes, the chatbot identified their eligibility, processed the refund, and confirmed the transaction—no hoops to jump through, no back-and-forth. The result? Not just a satisfied customer, but one who was so impressed by the ease of service that they considered resubscribing. This is the hidden power of great CX: when support is frictionless, it doesn’t just resolve issues—it builds trust and loyalty.

Human-centered AI is an approach that prioritizes people’s needs and values, aiming to make AI interactions more natural and user-friendly. There are plenty of ways to work with your AI to achieve this outcome. 

1. Ditch the Scripted Responses

Image Source: Convince and Convert

The fastest way for a customer to figure out that they are speaking to an AI agent versus a real person is getting a scripted response. Have you ever been on the phone, and the generated AI keeps repeating itself, and you end up frantically yelling “real person” over and over again, hoping to bypass the robot you are chatting with? Even worse when you’re online with an AI agent and like the example above, you keep getting the same scripted response that doesn’t acknowledge or authentically respond to your problem. 

Swap out the scripted responses for a conversational prompt that will train your AI to sound authentic and genuine instead of transactional. Human-centered AI prioritizes people’s needs and values, which can enhance their user experience.

Image Source: Tidio

When generating a conversational prompt, you focus on what the user hopes to achieve from the interaction instead of just executing based on predefined tasks.A scripted response can be technically correct, but it misses the context and dialogue that address the true needs of the user. 

This example from Tidio shows how despite short responses from the customer, the AI agent is still being conversational and genuine in its replies. From a quick glance, you can’t even tell that she’s interacting with a chatbot. The natural, engaged, and emotive responses are a great example of good customer service that is attentive to the needs of the customer. 

While there are several instances where a scripted response can still benefit the user, conversational prompts are user-friendly and support human-centered AI, which places the user’s well-being at the forefront of the conversation.  

2. Determine Your Tone of Voice

If your tone can be easily misconstrued when you’re DMing a person on Instagram or sending a Slack message to a co-worker, it’s safe to say that AI is going to get it wrong as well. Every brand or business has a unique tone of voice that reflects their point-of-view and expertise. 

I know that using my chatbot to renew my prescription at the pharmacy will sound different than telling my Alexa to order groceries (at least it should). If you’re trying to decide the best way to transition to human-centered AI, start by reviewing your tone of voice.

You never want your AI to generate content or responses that don’t align with your brand’s tone of voice, so you need to check that your AI system is trained in your writing style. 

According to Merriam-Webster, tone of voice can be defined as the “way a person is speaking to someone,” do you know how your AI sounds to a customer? Is it formal? Serious? Funny? If you don’t know this offhand, then it’s time to review your content library. 

Forethought makes it easy to elevate the quality of responses your AI agent generates with their “Tone of Voice in the Solve Widget.” In just a few minutes, you can briefly describe a tone of voice that matches your desired style. If you are drawing a blank and are unsure what words best describe your “tone,” then you can select from suggestions like “friendly,” “professional,” or “straightforward.” 

AI tools can be trained on tone of voice, but your content is what will help refine the responses. 

3. Personalize With Your Brand Guidelines 

Brand guidelines are often the cornerstone of any company because they help maintain consistency across blog posts, social media, pitch decks, and more. In the same way your brand guidelines help to create a clear picture of your company for prospective and existing customers, you can use them to help support human-centered AI for your chatbots. 

Your AI can use your brand guidelines to produce content that reflects your unique voice, tone, style, and positioning. We often use a roadmap for every product we plan to roll out each quarter; think of your brand guidelines as the same approach for your AI. Your brand guidelines can help determine the writing style that your generative AI uses, whether you prefer MLA or Chicago format, or even your target audience’s characteristics.

Image Source: Medium

Generative AI can get as granular as needed, but the devil is in the details. If you haven’t already determined how your brand guidelines interact with your AI, it’s time to tweak them so they can support your current AI tools more effectively. For a retailer like Sephora, they need their chatbot to replicate the same informative, fun, and conversational language that a customer would experience in the store. In our examples above, Sephora’s beauty chatbot combines personality and fun in how they address the customer while also providing beauty tips and product recommendations. 

It doesn’t get more up-close and personal than having your AI as the first touchpoint your customers have with your brand. Ensuring your AI reflects your brand’s personality and values can help create a cohesive user experience that sounds more human and less like a robot. 

4. Get Comfortable With AI Editing

Calling all copywriters, content marketers, and editors, it’s time to learn a new skill: AI editing. Human-centered AI can be a game-changer for enhancing a customer’s user experience because it makes it easier to understand and engage with. Editing your AI content is essential for improving your readability, specificity, and relatability to your customers. AI-written text can be factually and technically correct but isn’t always readable to a customer. 

When you take the time to edit your AI-generated text, you can maintain the authenticity of your content and build consumer trust. Here are a few tips on converting your AI text to human text.

  • Start using simple, compound, and complex sentences to improve the flow of your AI text.
  • Incorporate everyday slang or idioms to help your AI text sound more natural.
  • Add humor and emotions to your AI with jokes and feelings to make your text more engaging.
  • Identify any unnatural phrasing that wouldn’t sound or read authentically to a human.
  • Check your AI settings to determine if you can adjust the formality of the language. 

Unlike traditional editing, you’re not skimming for grammar; AI editing requires you to run a fine-tooth comb through your content to identify awkward phrases, sentences, or unnatural words. 

5. Empower Your Agents With Forethought Assist

We’ve discussed several hands-on approaches you can take to create human-centered AI that will have that perfect blend of expertise, authenticity, and trustworthiness to connect better with your customers. But if you’re looking for a tool that will save you time and do the legwork for you, look no further than Forethought’s Assist tool. When considering the right tool to help improve your AI, you want to opt for a product designed with not just any person in mind but your customers. CEO Deon Nicholas shares how his experience with customer service in high school laid the foundation for his approach to integrating it with AI. 

“Forethought became a company focused on using generative AI to transform the world of customer service…When it comes to customer service, the resolution rate is king. Are your customers getting their questions or concerns resolved? This is another area AI can help with, as it provides you with insights that help you improve your resolution rate.”

You can use Assist to deliver an exceptional customer experience with a quality conversation that isn’t cold and technical. Whether it’s working through help desk tickets, knowledge articles, or response generation, Forethought’s AI empowers your agents with quick, efficient, and customer-centric tools. 

Nicholas elaborates on the importance of human-centered AI by sharing that “tools that allow agents to use natural language and conversationally address customer issues are much more effective and empowering.”

Conclusion

As AI evolves, it continues to be a solution to helping ease the strain of everyday tasks whether you’re using it to improve your customer service experience or writing a research paper. Still, it’s important to remember that it’s not a real replacement for human writers. 

Humans have sentence structure and language that sound like everyday conversation, a quality that AI is still striving to achieve. This conversational aspect of communication is what makes human writers irreplaceable, especially when you just “want to talk to a real person.” 

Using AI tools like Forethought can help improve and support your customer service strategy whether you’re planning for peak season or testing out new ways to connect with customers. 

Dive in.

Interested in generative AI for customer support? Check out this guide to learn about the 3 key pillars you need to get started.

Get your copy
Decor Half Circle Orange
Call to action decor Call to action decor