Insights from loveholidays on deploying AI to turbocharge growth

Livingbridge invested in loveholidays, the UK’s largest and fastest growing online travel agent, in 2018. loveholidays has since expanded its product offering and successfully launched into European markets – operating in Ireland and Germany in addition to the UK – in part thanks to innovative and deft deployment of AI. 

The travel-tech company, managed by CEO Donat Rétif, employs over 350 people across offices in London and Düsseldorf and is on a mission to become Europe's number one package holiday provider. The online travel agent is on track to help more than 4 million people from the UK alone to go on holiday this year, and is the UK’s third largest ATOL holder.

Notably, in recent years loveholidays has also been ahead of the curve when it comes to AI and machine learning, leveraging the technology to turbocharge operations, enhance customer experiences, and deliver efficiencies.

The company’s LLM-powered chatbot, Sandy, was launched back in 2020 to help with a surge in customer requests during the pandemic, long before AI became a hot topic in the travel sector. Sandy has proven a hit with consumers and now answers 77% of all customer service queries - significantly increasing the efficiency of the loveholidays contact centre and saving customers time.

And that was just the start. Re-shaping operations around AI developments helped the company speed up its entry into Germany in 2023, with content teams using generative AI tools to rapidly create locally-specific content. Today, the company is innovating further, using DeepL AI for translation and rolling out AI voice interactions with customers.

Here, our Investment Director, Amy Hunter, sits down with loveholidays CTO, Mike Jones, to glean insights into effective approaches to AI and learn more about the company’s plans going forward.

Amy: You’ve already achieved a lot with technology at loveholidays. How did you approach integrating AI?

Mike: We've been thinking about machine learning and AI and how they can help our business for a long time. The way I see it, there is a two way relationship between technology and the structure of a business, so when I joined as CTO in 2019 the first thing we did was create cross-functional teams around business problems. This helped us brainstorm potential use cases for AI and make developments including creating machine learning-based pricing models.

Then, when Covid hit, we expanded our experimentation with AI. Our phone systems and our contact centre software and teams just couldn't cope with the level of demand for refunds. So we started playing around with chatbot technology to start answering people's questions, and that's where it really got started.

Amy: Can you give a few examples of where AI is having a positive impact on productivity and revenues at loveholidays today?

Mike: AI is having a huge impact on loads of different elements of the business - it’s not just limited to efficiencies in our customer service teams, which is the use case that people often discuss.

For example, AI has turbocharged productivity in our five-person content team. We integrated ChatGPT into Sanity, our content management system, and now the team is able to prompt the tool to generate descriptions for every hotel on our website. Team members then edit this output, which is a much faster process. Overall, integrating AI has helped our content editors become 80 times more efficient.

Before AI, our content teams were manually managing around 2,000 hotels, with the capacity to update descriptions for around 1,000 hotels each month. Now, with the help of AI, it takes just two days to replace content descriptions for over 35,000 hotels linked through our website. This time-saving has a positive impact on our bottom line – costs have fallen to around 5p per hotel description.

Other examples include using DeepL AI for translation, which has helped us reduce costs for translating partner content. We have also used generative AI to upscale images which resulted in a notable uptick in conversion and helped us save around 500 hours of human time.

Amy: How did you create an AI-receptive culture at loveholidays, and help the team get confident using AI tools?

Mike: For us, AI is designed to augment human interaction, not replace it. That's really important. The technology has not taken away any content team jobs – we could never have hired the number of people needed to create the level of content AI is helping us deliver today. AI is instead making the team's jobs better by automating boring processes, freeing them up to work on more interesting and useful content for customers. We see AI as a means to help people who understand the business become more efficient.

It has been helpful to commission training for our teams, and we recently completed a training initiative with Google that enhanced AI literacy. The fact that we have centred the team in our experiments with AI from the start has also helped with integration and confidence using new tools.

We are open to ideas from our brilliant tech team, and we have promoted people internally to help us develop our in-house tools. For example, our lead prompt engineer - the awesome person building dialogue flows for our chatbots and voice-activated AI - is called Juliana. Juliana was an experienced [customer] contact centre employee, and she is helping us build the new tools because she is the person who really understands how customers interact with us.

Amy: Are there key things businesses should consider doing to start integrating AI successfully, as loveholidays has done? For example, how do you ensure AI does not run away and become a "black box" for the business?

Mike: AI can give everyone a boost, and if you have the ability to exploit this technology and you don't, your company will miss out.

My advice to any business starting out with AI is to think about internal processes and how you can improve them with AI, before doing anything for customers. This minimises risks and maximises control over the technology’s impact on operations.

I would suggest conducting a comprehensive analysis of your business processes and looking for repetitive, rule-based tasks where AI can easily integrate. The next step is then to conduct pilot projects to experiment with AI applications, focusing on solving specific business challenges. It's important to note that you don’t need tech teams to do this – a key advantage of this technology is that it puts businesses in the driving seat. The latest version of Google's AI tool Gemini sits on top of data, so you can ask what were our sales last month? and it tells you.

In my experience, succeeding with AI also relies on encouraging a culture of experimentation. To create and scale in-house AI tools you need to operate more like a tech company and allow teams to experiment. After that, it's about having the structured systems in place to allow you to take advantage of AI developments.

When it comes to the worries around AI becoming a "black box" for the business, I think it's all about being structured in what data you expose to AI, and making sure you are careful and have the right policies in place.

Amy: How do you ensure that you stay on top of the latest developments?

Mike: Continuous learning is key. This is such an exciting space and it's evolving all the time, so executives - even non-tech focused executives - need to invest in staying informed about the latest AI trends. I go to events on AI all the time, and try to make sure that we sign up to all the relevant training programs.

I also believe in hiring people to have a big impact and ensure we stay at the top of our game. We now have a team of around 100 people across data, tech and product, and these are all talented people who can build great things within organisations given the right autonomy and support. They bring in a huge amount of knowledge and help us ensure we don't miss major steps forward.

Amy: Finally, what is your vision for AI and loveholidays going forward?

Mike: The recent development that I’m most excited about is our AI voice interactions. We’ve been exploring these with Google, and around 25% of all voice interactions with our customers are now automated by AI. This makes customer conversations more natural and personalised, and we believe it will have a big impact long term, as using the technology is set to deliver £3 million in costs saved this coming year.

We’re now experimenting with personalised voices for our virtual assistant too, and it’s getting exciting. We're working closely with Google to do things like improving pauses in conversations, and using customer relations experts in our contact centre to help make it as relevant and personal as possible.

Overall, we're planning to leverage AI to help us continue pursuing our international expansion and growth. AI is a tool that is allowing us to move faster with better decision making and increased efficiency, and to achieve our goals in a much more profitable way.