How to avoid vanilla GPT

There's an interesting trend that often happens with technology and innovation. Something new and exciting hits the market – something offering near-infinite opportunities for innovation and progress. But then applications for it can seem rather… dull.

The technology behind OpenAI’s GPT models is a great example. As miraculous as it is, and for the impressive and creative ways some have used it, have most brands done anything truly impressive?

In many ways, it’s not a technical limitation but a human one – our creativity sometimes struggles to live up to the advances of modern tech. We’ve created incredible multifunctional tools, but we’re still using them like hammers.

And for many (but not all) brands, they’re struggling to use technologies like this to advance their brand, introduce repeatable and scalable efficiencies, and offer a true point of differentiation. Some have even tried to build their own large language models (LLMs). Which is like inventing a new language and dictionary, when one already exists. Just because you can doesn’t mean you should.

So let’s get the creative juices flowing. Because there are a lot of ways to utilize ChatGPT and other LLMs in business, commerce, education, government, and plenty of other parts of life.

Let’s add a little flavor to Vanilla GPT.

A new human interface for LLMs

A quick glance at what brands are using LLMs for today suggests room for a little (or a lot) added flavor. Between the sloppy but improving AI imagery (six-fingered humans, anyone?) and an overabundance of generic copy, it's fair to say that many businesses haven't quite hit their stride with generative AI yet.

It's okay, we get it. Enterprises know they need to be ahead of the game with LLMs, but they also need to make sure the output resonates, is ethical, and is safe for customers. It's a delicate balancing act and one that can’t be rushed into.

You might know that at UneeQ we make digital humans. Essentially, these are modern interfaces for AI systems like ChatGPT, Anthropic’s Claude, Gemini, Grok, etc. While LLMs create content and can reason like a real assistant, digital humans take it a step further, allowing users to interact with AI as if it’s a real human – face-to-face, in real time.

Synchronize some specific data – brand guidelines, product catalogs, CRM platforms – using AI orchestration tools like Synapse and you’ve got an AI that’s trained on your brand and ready to express themselves with precision and personality.

No more vanilla AI. Here are some ways to take your LLM to the next level. 

1. Turn chats into real conversations

One of the downsides of traditional chatbots is that they're one-dimensional. Loaded with predefined answers and limited to text-based exchanges, many chatbots are only useful for simple tasks or steering customers toward a human rep. 

LLMs go a step further. They can provide interactive, realistic responses that sound much closer to a genuine conversation. Close, but it's not quite like the real thing. 

What's missing? Well, a conversation is more than just about words. When we talk to another person, we engage with their voice, tone, and expressions. We interpret their gestures, body language, and movement. We empathize with their feelings and laugh at their jokes. 

Without all this, a conversation lacks soul – it's just a script. 

Digital humans can provide these essential elements, elevating LLM 'chats' into real conversations by bringing a voice to the words. And with our AI animation system, Synanim, we can make digital humans – and by extension, your brand – feel more alive through realistic, organic movements and expressive emotions.

It’s like taking that script and hiring an actor to bring it to life in more ways than just two-dimensional content.

The results are interactions that feel much more natural, leading to deeper user connections and customer experiences that are fun, educational, and engaging. 

2. Provide more meaningful insights

Getting information is one of the main reasons people speak to ChatGPT and other LLMs, according to Reuters Institute research. 

In fact, nearly a quarter (24%) of the general public have relied on generative AI for this purpose, only slightly less than those who have utilized it for creating media (26%). More specifically, 11% want ChatGPT to answer factual questions, while 10% turn to it for advice.

Unfortunately, the same study shows that some people are still unsure of whether LLMs tell the truth. Of those who have used generative AI to obtain factual answers, less than two-thirds say they trust the results. 

Some of this skepticism may be well-placed – a topic we've covered before on the UneeQ blog. However, even when ChatGPT is completely accurate, some of its responses can be overly broad and run-of-the-mill.

This is the job of Synapse. Synapse uses RAG to help digital humans stay on-brand and help deliver more accuracy and relevancy, well beyond what an LLM alone can offer.

If you take our previous analogy of having the script (LLM), and the actor (Synanim), Synapse is the director who keeps everything connected and focused. And it allows for more meaningful insights beyond what a generic LLM might offer.

3. Bring the AI personality to life

Synapse not only makes brands more comfortable using AI for engaging customer-facing applications, but it also allows them to build a personality for the front end.

Just look at Emma from The City of Amarillo. She has a backstory, a set of values, and a personality type built into her model, so all responses have a certain flair only Emma can deliver.

This goes for all digital humans. If I were to ask you to describe the personality of ChatGPT, you’d struggle.

But I’m sure you’d be able to pick out the personalities of digital assistants like Sama from Qatar Airways, Emma from Amarillo, Andie from Dell, or Max from Deutsche Telekom.

The personality is determined in the digital human creation process and expressed every time someone has a conversation with the AI.

4. Let users take action 

ChatGPT is a great conversationalist, but its functionality beyond that is relatively limited. So, what if your customers want a service that walks the walk, as well as talks the talk?

Digital humans can be more than just a confidante who provides useful information and advice to customers; they can act as a concierge who takes actions on a user's behalf. 

Need help navigating a website? A digital human has it covered. Want specific content, such as videos or images brought up on screen? No problem. And if a client or customer needs to speak to a real person, a digital human can escalate their query, ensuring the user's details are jotted down for a follow-up.

Some of our AI assistants have even transacted with customers directly from their kiosks.

Whereas most conversations with ChatGPT just end, digital humans are equipped to keep the interaction going. We would say they offer 'less talk and more action', but it's actually 'more talk and more action'! 

The upshot? Your company can provide richer, multifaceted experiences that engage people throughout their customer lifecycle.

To discover more about how digital humans can add some flavor to your vanilla GPT offerings, book a meeting to discuss with one of our specialists.