Seven key insights from creating an award-winning AI chatbot

This article was written by Benjamin McCulloch, Content and Conversation Designer at VUX World.

AI assistants can work wonders in all sorts of industries. You may have been put off by overhyped and under-delivering bots in the past, however there is now concrete proof that AI assistants can deliver huge returns on their investment.?

We have the case studies to prove it, and this one’s a whopper! It’s an award-winning chatbot for a debt recovery client, created by VUX World.?

It’s been successful for many reasons. Here are 7 of them that’ll put you on the path to AI success!

If you thought you needed to be a developer to build an AI solution, think again. Learn from Bret Ingerman from Tallahassee State College how it’s benefitting from AI automation and from Cobus Greyling on how you can build your own AI chatbot for your business without the need to code.

On August 1, we're hosting an exclusive webinar to dive into the key benefits of implementing AI chatbots and how you can identify best use cases that work for your business and your customers.

Discover more

1. Know what you aim to achieve

Before anything gets built, you need to know what you want to achieve. You will save time, money and the embarrassment of releasing a ‘dumb bot’.

The most important step you can make when creating a conversational assistant is to research what people need from it. What are common things people need from you that are simple to provide?

That’s what we did with our debt recovery client. We analysed transcripts of conversations with their customers to find the most common needs. As it turned out, most people wanted to find out how much they should pay, when they should pay, and how they could pay.

We discussed ways that those needs could be resolved. It’s often a case of checking what live agents do, and finding the best way to automate and improve those conversations. That’s where you should start. From that solid foundation you can continue to improve your assistant.

2. Connect your assistant with the right data

To be useful, an AI assistant must know things. Think of it like a new employee on your team. What would they need to know to do their job??

Once you have defined the required information, you need to connect the assistant with those data resources.?

This can often be easier said than done. In this project, there was a few month gap between Discovery and Alpha while we worked with the client’s systems provider to obtain the relevant APIs needed to bring about the use case.?

3. Have an open mind

People relate to AI assistants in surprising ways. For example, studies have proven that people can be more comfortable sharing personal information with an AI assistant than a person.

Debt recovery often involves uncomfortable conversations. People may feel like they’re overwhelmed by debt and can’t imagine how to pay them off. They also don’t want to feel judged for being in that situation. A carefully designed AI assistant can help them to open up.?

It’s vital that the things they share are secure and private, not just from a legal and ethics standpoint, but also because they need to trust the system to feel comfortable disclosing personal details with it.?

This is why we worked on the assistant’s personality to make sure it is open, non-judgemental and personable, just like the agents are. (FYI, we used Rebecca Evanhoe and Diana Diebel’s Personality Canvas from the book Conversations with Things)

4. Act on learnings

While we’re on the subject of keeping an open mind, you should expect to be surprised by the conversations your assistant has with people too.?

Choices made during the design phases should always prioritise the user experience, but the users will still surprise you! We’re talking about real people living out their lives in a complex world, and your AI Assistant must be an active participant in that user’s unique world.?

People will say things that you don’t expect. They will expect things you didn’t imagine. That’s fine, and a normal part of the process! You just need to pay attention, and consider how to improve the experience based on the things you learn when real people interact with it.

At VUX World, we always say that going live is the beginning of the journey, not the end.?

5. Keep one eye on the big picture and the other on tiny details

How do you know if your AI Assistant is performing? It needs to be monitored continuously. The individual conversations need regular transcript reviews. That gives you an idea of the experiences customers are having.?

You can’t read every conversation, though. VUX World created a robust analytics suite to monitor the performance of every moment in conversations to ensure they’re all working as they should, all the time. This is how our client learned that their chatbot had helped move ￡1 million+ to payment arrangements within a year. Talk about ROI!

6. Create a roadmap

A great AI Assistant is going to transform your business. It will improve customer experience while also strengthening your organisation.?

Once you’ve deployed your assistant and it’s achieving its initial goals, you should plan your roadmap for continuous improvement.

For our Debt Recovery client, they started looking for other ways to infuse CAI in their business to take the strain off their live agents. Ultimately, they found that they needed one less full-time agent (saving them ￡27,500 per year) and the remaining live agents help customers with complex needs.

To this day, we have a roadmap that we groom and implement each month.

7. Bring the assistant into your business pitch

You should think of the AI Assistant as performing a crucial role in customer service. It’s a front-line representative that should help users get what they need faster than a human could serve them, and that means your human customer service agents can be laser-focused on the customers who have complex needs.?

You should also look at it from a business perspective; you now have an AI system that was purpose-designed for your needs. It allows you to scale up as needs arise, because the AI assistant is available 24 hours a day, 365 days a year.

You’re succeeding with AI, and let’s face it, we’ve all read headlines about people who aren’t doing AI well! Since launch, our client’s chatbot improved CSAT by 30% and reduced query resolution times to 2 ? minutes (saving 400 hours of agent’s time in a year). They take these figures into every sales pitch.

When you get your own successes, you definitely should bring them into your sales pitches to prove that your team is enhanced by AI, and that it can provide improved customer service to untold numbers of customers.

The chatbot VUX World created for this client won the CIVEA Innovation in Enforcement Award 2024. You can download the full case study here or reach out if you’d like to discuss how we can help you transform into a successful AI-infused business too.

To dive into this case study in more detail, you can rewatch the webinar we recently ran, or you can download the case study here.

About Kane Simms

Kane Simms is the front door to the world of AI-powered customer experience, helping business leaders and teams understand why AI technologies are revolutionising the way businesses operate.

He's a Harvard Business Review-published thought-leader, a LinkedIn 'Top Voice' for both Artificial Intelligence and Customer Experience, who helps executives formulate the future of customer experience ad business automation strategies.

His consultancy, VUX World, helps businesses formulate business improvement strategies, through designing, building and implementing revolutionary products and services built on AI technologies.

1. Subscribe to VUX WORLD newsletter
2. Listen to the VUX World podcast on Apple, Spotify or wherever you get your podcasts
3. Take our free conversational AI maturity assessment