Steve Endacott argues travel AI implementations are set for take off
Like many of us, I became fascinated with the potential of generative AI and large language models (LLMs) when the first versions of ChatGPT were released.
However, becoming chairman of AI Incubator Neural River has created a day-to-day immersion in the AI space for me and a focus on how to create practical use cases in travel and other business sectors.
I would like to puncture the hype and dispel many of the myths about Gen AI and focus on practical advice for using AI in the travel sector.
The biggest weakness of ChatGPT-3.5, launched by OpenAI in November 2022, was that the model was only trained on data up to September 2021 and knew nothing about the world since this date. So, although we could see the future of AI, it was hard to implement.
On November 6 last year, OpenAI launched ChatGPT-4 Turbo which feels like a watershed moment as it is built on an underlying database with a knowledge of the world up to April 2023 and with links to the Microsoft Bing search engine for up-to-the-minute information.
Initially, this version was made available only to developers and businesses via the GPT API, but it will be opened up to consumers.
In any AI application, the quality of foundational information is critical, and this new version delivers the necessary high-quality data. Though it might not appear as an exciting development, this advancement ensures a solid, reliable base for AI applications, essential for their successful deployment.
However, it is the increase in the ability to shape and control ChatGPT outputs from 3,000 words of input instructions to 300 pages that is likely to have the biggest impact.
Additionally, with the launch, Open AI is encouraging AI developers to create applications and post them as chargeable plugins in their equivalent of an app store.
The diversity of products is expected to be vast, from tools that enable automated analysis of a company’s pay-per-click results to algorithms for optimising bidding campaigns and tools to automatically increase prices on products selling ahead of expected yield curves based on historical sales patterns.
Monitoring the GPT marketplace and picking up pre-built tools needs to be a top priority for any business IT team.
Formulating an AI roadmap for your business is crucial. It should enable strategic prioritisation in the allocation of internal resources and guide the implementation of AI tools within your organisational framework.
While the allure of AI lies in innovation and pioneering new technologies ahead of competitors, the true value of AI is realised through its application in enhancing day-to-day operational efficiency. Often, the most effective tools for these purposes are available as off-the-shelf solutions.
If I was still running a travel business, my top three AI priorities would be:
- Content creation: utilising ChatGPT as a content co-pilot can enable travel businesses to generate extensive databases of customer-centric content, including detailed information about resorts, destinations and local amenities. Such content can be tailored to suit the preferences of different groups, offering personalised recommendations for bars, restaurants, and attractions. This enhanced level of customisation is vital for travel agents and online travel agencies (OTAs), especially those without representatives in resort. However, caution is advised against over-reliance on AI for SEO (search optimisation) purposes. While AI-generated content can initially boost search engine rankings, this may lead to transient benefits and potentially incur penalties from search engines such as Google, adversely affecting your site’s credibility.
- Customer service improvements: 80% of customer enquires can be dealt with using a database of standardised answers either delivered by people or increasingly by AI Voice Agents, provided by businesses like my own neural-voice.ai. Just have a play to see the power of these tools.
- Yield management and Google bidding: both these crucial aspects hinge on analysis of large datasets, a domain where AI excels. AI’s capability to process and interpret vast amounts of data makes it an invaluable tool in yield management and optimising Google ad-bidding strategies. By leveraging AI, businesses can make more informed, data-driven decisions that enhance efficiency and profitability in these areas.
It’s essential to dispel prevalent myths about AI. For instance, while AI can produce inaccurate or fabricated responses, often termed ‘hallucinations’, this issue is significantly mitigated with proper training on relevant data. Additionally, concerns about GDPR compliance are often unfounded, as platforms like OpenAI do not record customer data, eliminating compliance issues.
These concerns are frequently rooted in a lack of understanding of AI technology, particularly among IT teams. This underscores the need for a shift in mindset and, potentially, personnel.
The AI revolution is firmly on its way so make sure you don’t get left behind or, worse, dismiss it as this week’s hot topic that will soon fade way. It most definitely won’t.
Steve Endacott is chairman of AI incubator Neural River