“Large language models are like a detective story,” said Matthew Newton, CWT’s VP and Chief Architect on a Business Travel Show panel in June 2024, “Gen AI allows you to predict the whodunnit.”
Indeed, it’s Gen AI’s ability to predict an outcome with accuracy that appeals to the masses at the moment. The future impact of AI is a leading topic of conversation on tech panels and at dinner tables the world over. Given the eye-popping speed, hype and investment that has gone into developing AI, some believe we’ve already become disillusioned. Market research firm Gartner said in its 2023 AI Hype Cycle paper that generative AI was at the “peak of inflated expectations” and fast reaching the “trough of disillusionment.”
Karen Hutchings, founder of Cobb and Hutch Consulting tells in BTN & CWT’s 2024 white paper, Emerging Technologies in Corporate Travel, “Technology is coming like a freight train. Traditional companies are not waking up fast enough.”
CWT’s enterprise architecture team develop digital strategies, manage enterprise technology environments, and deliver customer-centric digital solutions for clients in 139 countries, to modernize core systems and enhance differentiation.
Seeing as a good prediction relies on the quality of underlying information, we ask the team for their view on the hype and reality of Gen AI in business travel:
There will be waves
"Gen AI won’t develop along a single curve. We’re seeing wave after wave after wave as new parties and new entrants enter the conversation. A trough of disillusionment is inevitable. When smartphones first arrived on the scene, it was the battle between that and the Blackberry. It was a similar story with the internet before that. The interesting difference this time around is the speed at which Open AI has become successful and the conversations that people are having are at every level and every function across the business. This isn't something that's just being spoken about in ‘technology.’ Gen AI brings together two facets of an enterprise, the business community and the technology community. That's always going to be a good thing."
- Matthew Newton VP and Chief Architect
Speed to market keeps Gen-AI relevant
"Sometimes people get excited about a hype and then disappointed when it doesn't play out either to their expectations or timescale. With AI we've seen real results in very short timescales and we can get an AI product to the market place far quicker than most technologies. We’re able to deliver interesting results for our customers, fast."
- Gordon Coale Senior Director, Enterprise Architecture .
Tried and tested wins the race
"Open API is accessible almost to everyone, which is very different from previous machine learning algorithms, large language models, or even transformers. Like many experiments, out of 100 ideas, maybe 10% will be successful. Like in many organizations, our product team experiments before converging around one or two ideas. The hype phase stabilizes but this phase is where the value lies."
- Xiaolan Sha Senior Director, Data Science & Machine Learning
Data with destiny
"When you have a large volume of data, you can use large language models to answer the right questions. You can use it to make agents more productive across varied clients and client policies. But large language models need to be trained correctly. They are probabilistic models trained by asking them to complete a piece of text. They will always try to respond. They don't tend to say ‘I don't know the answer.’ Users of AI need to apply safeguards, employ domain engineering and a robust validation framework to avoid hallucination. Models need to be scored and fine-tuned but the impact on productivity is enormous."
- Leonidas Constantinou, Senior machine learning engineer
Whether related to cost-savings, personalization, or security, the potential for Gen-AI to boost productivity and combine disparate data to deliver insights at speed, its potential is vast. To find out more about the opportunities, risks and challenges of Gen AI and other applications, read Emerging Technologies in Corporate Travel.
