TUI breaks down their ai approach step by step

Plus Google releases its Gemini model

We are just 23 subscribers away from 1500. That’s pretty cool. Hi everyone! 👋 Room for loads more if you know anyone who could benefit.

Last week Amazon, this week GOOGLE

Not a week goes by now without a new release of an ai model into the ecosystem. Everyone is scrambling to ensure that OpenAI don’t open up an unassailable lead with their GPT products. Of course OpenAI itself was nearly left dead on the floor 2 weeks ago!

This week Google released its Gemini model and the demo video showing its multimodal capabitilites around image and voice are pretty darn amazing.

Of course demo videos are supposed to be amazing so I’m looking forward to having a play myself and see how it shapes up.

It is always perhaps just worth waiting a week or two post launch, before placing your bet on any of these emerging technology platforms. This week Amazon Q was reported to be “experiencing severe hallucinations and leaking confidential data”. Ouch.

TUI breaks it down

What I love most about this journey through ai is that everyday you are learning. Once you think you are getting a handle on something, more information gets shared adding a new layer to the knowledge bank.

At the (seemingly) huge Amazon AWS event this week was a specific industry breakout featuring TUI and Manchester Airports. This video of the presentation on Youtube is remarkable viewing. Actually if you are a CEO or CIO, you should probably stop reading now and just go watch the first 30 minutes of this video.

The TUI priorities

It was an incredibly honest share by Mark Jennings, CIO of Analytics and Artificial Intelligence at TUI who takes us through the past 8 months inside their business, the opportunities they’ve decided to tackle, the tech stack they are using and even exactly how they are putting the pieces together.

It isn’t a watch for the casual observer TBH, but if you are building internally or strategising around your build then it is compulsory viewing IMO.

With my HandbookFM.com founder hat on, it was pretty exciting to see one of thier 6 areas of priority was to “accelerate training & on-boarding” as that is exactly the tech we’ve built. As HandbookFM utilises all existing documentation a business already has & takes about 90 seconds to produce the podcast style audio training per policy or procedure, that automated training & onboarding could start tomorrow at TUI. Maybe we should chat? 😀 Call me!! 📞. Anyone else can also try it out for themselves cost and obligation free over at HandbookFM.com 

If you think someone (or everyone) you know or work with could grow from being more informed on the topic of ai + travel (or could use the training above) then please forward this email to them and they can click the button below:

South Korea’s metro is using ai for real time translation with tourists

A very cool case study here of the Seoul Metro takes us into the experience of live translation for metro customers. The system they have built has both hardware (microphones, speakers and some interactive OLED screens) and software components doing the translations.

The are currently supporting 13 languages including English, Japanese, Chinese, Vietnamese, Thai, Malay, Bahasa Indonesia, Spanish, French, German, Arabic and Russian and in one major station for the pilot.

Translation is one of the things that was previously very hard and is now relatively simple. Expect a flood of companies in 2024 to look at expanding their markets using real time translation of voice and content translation for their acquisition.

Speaking to one leader at a major experiences player this week, they weren’t sure if the guide jokes would translate well. Good point!! Wonder if the transport staff in Seoul are cracking any funnies to test this out?

Got a tip or seen a story I’ve missed? Let me know by simply replying to this newsletter.

Radisson takes the ai route to win RFP’s

Fascinating story this week around the hotel chain Radisson introducing ai tech into their request for proposal process when it comes to pitching for meeting and events.

I love this usage because this has surely come from someone who deeply understands the pain and inefficiency in the current process and has spied the opportunity that ai presents to fix it. Radisson have partnered with German company Hivr.ai in this endeavor.

So does it work? Absolutely it does! “Radisson said that from late last year to September 2023, it increased its lead to quote conversion by 80 per cent and doubled its customer response rate. The technology also reduced the time needed to send proposals by 59 per cent.” Those are some numbers!

The tech helps “improve response rates for buyers and saves time on manually processing requests.” It is these types of innovations that will see the early movers take market share whilst their competitors are left scratching their heads wondering what has happened.

What are your manual processes that gain you an advantage? Go ask you team! Does it take 2 hours to put a customer itinerary together, copying and pasting from loads of sources into a word doc which you will then PDF? Even the non ai software in this space is a laborious manual process to build each itinerary. Is someone spending their day trying to format data? Are your sales team clogged answering basic information questions? The answers are out there!

As Carlos Munoz from Hotelbeds said this week at MarketHub Asia ““t the moment, there are many reasons why it is a painful, laborious process to create a travel experience for the customer. AI can remove those pain points and help create the right itinerary.”

Not sure where to look? Reply to this newsletter and I’ll point you in the right direction.

Airports are rapidly changing with ai

Toronto’s main airport is the latest to reveal ai fueled innovation to help us get through the airport and out to the destination faster.

In this example they are applying vision ai alongside their existing set of cameras in the immigration area and send real time updates, which can result in officials opening up new lanes and help clear backlogs. They report average waiting times have reduced “from an estimated 30 minutes during peak periods in 2022 to just under six minutes last summer.”

Toronto’s Pearson International Airport has partnered with Zensors AI platform. Anuraag Jain, Zenzors cofounder and head of product and technology said ““We provide a platform that allows airports to instead think more like software companies, deploying quicker, cheaper and more accurate solutions using their existing cameras and the latest AI technologies” <insert raised eyebrow emoji>

Unfortunately many heading into the airport will likely be flying Air Canada and that is a pain that ai is unlikely to ever be able to eleviate.

Small Vacation Rental agencies being left behind

An article reporting on a Hostaway report showed “Just 34% of managers with under 10 homes state that they are using AI in some form in their business, compared to 53% of managers with more than 50 properties using AI tools regularly”.

Hostaway does have a dog in the fight however. As the article explained they’ve “launched a ChatGPT-powered artificial intelligence tool within its property management software capabilities to support vacation rental business customers to manage listings and increase bookings, by perfecting listings on the OTAs. The tool matches the tone of voice and professional style to each operator’s brand and audience to create brand consistency and can be used no matter what the size of the portfolio.”

Actually I thought 34% was pretty high based on what I see & hear across other parts of the industry.

Where are the women?

Not specifically tourism related but I felt this was important to help shed some light on. This Linked In post by Daniela (Dani) Herrera called out a New York Times story on where they “published a "Who’s Who Behind the Dawn of the Modern Artificial Intelligence Movement."

And the number of women on the list? 5? 2? 1?

Nope - they found zero. None.

Herrera however found many:

🔹 Joy Buolamwini: Best-selling author of Unmasking AI, Founder of Algorithmic Justice League, and Researcher. Her methodology uncovered large racial and gender bias in AI services from companies like Microsoft, IBM, and Amazon.

🔹 Daniela Braga: CEO of Defined.AI, the largest marketplace of ethically sourced training data for AI.

🔹 Kieran Snyder: Co-founder and CEO of Textio, a language analysis platform that uses machine learning and natural language processing to help organizations create more effective and inclusive communications.

🔹 Timnit Gibru: founder and executive director of the Distributed Artificial Intelligence Research Institute. She was one of the Ethics in AI Researchers fired by Google after she raised concerns about their technology.

🔹 Margaret Mitchell: A senior AI Ethics Researcher "Advancing AI informed by human values.". Google also fired her after raising ethical flags about their tech.

🔹 Fei-Fei Li: Co-founder and chairperson AI4ALL, a non-profit focused on increasing inclusion and diversity in AI education. She's also the inventor of the dataset and benchmarks that contributed to many of the AI developments we know.

🔹 Rana el Kaliouby: Co-founder and CEO of Affectiva. She's a pioneer in the field of Emotion AI.

Let’s do better people. Be conscious to your unconscious bias.

One for the nerds

This was my second favourite discovery of the week after the TUI video above. This one is a little deeper in the weeds but hopefully helpful for those tinkering away building their own small solutions. But first - an anecdote!

A couple of decades ago I was training to be a driver guide in the Australian Outback. I’d made my way through the company training on how to survive with 25 backpackers out in the wild but hadn’t yet got my licence to drive a larger sized bus. Due a string of unfortunately circumstances there never seemed to be a vehicle around to go out and get some miles under my belt until just the evening before the test. One of the trainers took me out for 45 minutes around town. It didn’t go awesomely well. He told me I’ve only got one chance. He told me to say, out loud everything I was doing, as I was doing it through the test itself and explain my process. I passed and the rest is a 25 year career in tourism.

The article I read this week was around a prompting technique with Large Language Models called Chain of Thought (CoT). The idea is to not just ask the LLM the answer to a question but rather to get it to explain how it is getting to its answer. And do this repetitively over and over. When it gives a poor methodology, instruct it to never follow that path again. Brilliant!

The article has a lot more detail………..

Slack Group!

Some cool discoveries in the Slack Group this week including Hoodmaps which shows your city but with some very funny and tongue in cheek, crowd sourced phrases of what each area of the city is about & also the back story of WhatsApp which started as a travel related idea!

Want in on that - sure thing - its free (for now and always will be for the early adopters, but maybe not forever) The moment you move from passively absorbing ai news in travel to actively doing anything at all - you need to be in this group. That is where real value is being created. (Sometimes this link doesn’t work - just reply to this email if you hit any snags)

This box used to be about new ideas. Those are now discussed in the Slack Group.

Can ai save coral reefs?

Flicking through Dive Magazine this week, it was great to see an article around how ai is playing its part in trying to save our threatened coral reefs. If you weren’t aware “scientists (are) estimating that a 1.5C increase in water temperatures would see 70-90 per cent of the world’s coral reefs destroyed.” Maybe there is something we can about that?

In the meantime “in the Abrolhos Islands, 60km (40 miles) off the coast of Western Australia an innovative new project is underway to restore coral reefs using a mixture of robots and AI technology. Small bits of coral are grafted into special limestone bases which are then planted on the sea floor.”

The work however is very difficult and slow to do by hand so “robot arms are being trained through AI to do the task instead of people.”

Note the trend here. Slow and laborious manual tasks. Please go ask your team what these are in your business if you don’t already know. They could be gone by February.

The number here caught my attention - 1 each week!

The travel planning war has been won! Get Your Guide have hailed a “revolution in the discovery of travel experiences!”

On closer inspection however it looks just like a list of things to do across 3 days in Paris. There is no interactivity. No place to add your preferences or simply type or speak the changes you would like made. There isn’t even a bot typing away to generate the itinerary in front of your eyes.

The “incredible prototype…fully built by Artificial Intelligence” is actually just a bit of broadcast content. The power of ai at the consumer level is to provide the chance for collaboration and input (you know, customer preferences) as part of the UX. Not just scream “buy my thing, its ai”. Unfortunately it just smacks of a business dying to get into the AI conversation anyway they can. Makes you wonder what is behind this? Potentially it is boardroom pressure to get in the game?

So game back on……. and then………. Paris Hilton invested in a travel planning startup

How to work with Tony:

Not exactly working with me but this week we have launched Ale Blazer into the Melbourne market. Ale Blazer is a member based community of craft beer lovers who want to support a vibrant and thriving craft beer scene in the city. Members get either 12 beers for $10 across the year on the one-a-month subscription or 52 beers for $40 across the year on the once-a-week plan! Who’s keen for a beer? We are pouring now. (Also the best Kris Kringle you’ll find!)

Got a question about ai? Ask it in the Slack group. I will probably give you my answer but you will also likely get 5-6 other opinions too.

Got an ai SaaS product or Tool: Sponsor this newsletter. Get your product in front of the decision makers. It is A$350 (but that’s Australian money so basically nothing if you live elsewhere).

Project Work: I work with you and your team on a specifc problem or opportunity. Depending on scope that generally is a 4 week process including a session with the stakeholders each week. The deliverable comes to you in a project board as a living document and roadmap to continue to reference through execution phases + setting up the first next experiments to test chosen hypothesis. My base rate is $4000AUD (+10% GST if you are in Australia) for a project that has a 30 day timeframe that includes the 4 sessions with team and then work outside of that to research and provide the recommendations, connections and introductions required. I’m now booked up until mid January but happy to take new year enquiries.

Become a customer of HandbookFM - workforce training automated. Turn policies into podcasts for simple oboarding and systematic training of policy and SOP’s (travel use cases here for onboarding/training DMC’s or training staff at hotels or airlines etc.

Most clicked last week was the X link to the 15 free courses to learn the basics of ai! The holidays break is a great time to delve into one or two of these and level up your game. We might also do another round of our courses with Sam Keller in early 2024 if you prefer the TL;DR, 1 hour version.

That’s it - you’ve made it to the end of this edition. I’ll be putting the result of the most clicked post in next week’s edition so you can see where others are focussing. If I’ve missed something, you’ve got a tip or any feedback at all - you can simply reply to this email and it will come straight to me. I’m doing this for You so please don’t be shy to tell me what you think

Glossary

Artificial Intelligence (AI) Artificial intelligence leverages computers and machines to mimic the problem-solving and decision-making capabilities of the human mind. (source IBM)

Generative AI (GAI) is a type of AI powered by machine learning (ML) models that are trained on vast amounts of data and are used to produce new content, such as photos, text, code, images, and 3D renderings. (Source Amazon)

Large Language Model (LLM) is a specialized type of artificial intelligence (AI) that has been trained on vast amounts of text to understand existing content and generate original content.

ChatGPT - Open AI’s LLM; sometimes referred to by its series number GPT3; GPT3.5 or GPT4. These are used by Microsoft & Bing.

BERT - Google’s suite of LLM. BARD is the most common of these.

If wanting to go even deeper into the AI lexicon - check out this handy guide created by Peter Syme for the tours & activity sector