Unlocking the power of LLMs: putting people first in AI assistant development

It’s easy to approach every new problem in conversational AI as if you’re the first person who ever faced it. But the truth is that many organisations are trying to solve the same problems at the same time.

For example, most banks offer mortgages. You can expect a banking AI assistant to be asked “when’s my next mortgage payment due?”

Yet, if you were to create an AI assistant from scratch to deliver on that use case, then you might end-up starting at square one: gathering a whole bunch of training data, then iterating and fine tuning your model until it reaches acceptable performance.

Domain-specific NLU systems, combined with large language models, present an opportunity to shortcut that training exercise, and expedite the time-to-value for companies with commonly shared use cases.

Raj Koneru , Kore.ai CEO and?Prasanna Arikala, CTO,?explained how it works during our recent?VUX World webinar.

Unparsed - the world's first Conversation Design Conference

We're?bringing?together the world's leading experts in conversation design to help you design more effective interactions with speech, conversational and NLP-infused interfaces.?Join us in London or online on July 24th and 25th.

Tickets are buy one get one free right now, so grab them while you can!

????Unparsedconf.com????

To get the offer, just add two tickets of the same type to your cart.

This approach could save developers and designers time. As Raj Koneru says, “the amount of work gets cut down anywhere between 30% to 70%.”

20 ways of using LLMs smartly

One of the hardest challenges every team has is obtaining good data.?You need training data for your NLU. It needs to be diverse so that it reflects the various ways customers will talk to your brand, and you need a lot of it.

A simple but effective use case is to us LLMs (Large Language Models) to generate a variety of training phrases. That allows the system to be trained incredibly quickly.

Once the bot is live, you can collect real examples from actual users to improve the data.

We covered 19 more examples of how large language models can impact your intelligent automation strategy in the webinar. You can watch the replay here or download the in-depth white paper below.

This post was sponsored by?Kore AI, and written by?Benjamin McCulloch. Ben is a freelance conversation designer and an expert in audio production. He has a decade of experience crafting natural sounding dialogue: recording, editing and directing voice talent in the studio. Some of his work includes dialogue editing for Philips’ ‘Breathless Choir’ series of commercials, a Cannes Pharma Grand-Prix winner; leading teams in localizing voices for Fortune 100 clients like Microsoft, as well as sound design and music composition for video games and film.

---

About Kane Simms

Kane Simms is the front door to the world of AI-powered customer experience, helping business leaders and teams understand why voice, conversational AI and NLP technologies are revolutionising customer experience and business transformation.

He's a Harvard Business Review-published thought-leader, a top?'voice AI influencer'?(Voicebot and SoundHound), who helps executives formulate the future of customer experience strategies, and guides teams in designing, building and implementing revolutionary products and services built on emerging AI and NLP technologies.

• Subscribe to my newsletter
• Listen to the?VUX World podcast?on?Apple,?Spotify?or wherever you get your podcasts
• Attend Unparsed?- the world's first Conversation Design Conference.
• Take our free conversational AI maturity assessment