Amazon enters the (ai) room

This week Amazon launched its new ai tool, Amazon Q (which is different to Q* which is covered below). Unlike the B2C focus of OpenAI, Amazon seem to be clearly targeting business & enterprise. “Amazon Q—a new type of generative AI assistant that is specifically for work and can be tailored to your business..” was how the live blog of the launch event described it.

Examples given were contact & call centres “assisting agents with proposed responses, suggested actions, and links to relevant articles based on what a customer is saying in the moment”.

As seemingly always, travel was centre stage, this time in the form of a car rental agency where “Amazon Q could detect a customer is contacting your rental car company to change their reservation. It would then generate a response you could send, detailing the company’s change policies and guide you through the step-by-step process of updating the reservation.”

In a nod to the recent GPT creator over at OpenAI, the launch also talked about the ability of managers & administrators to “create even smarter chatbots for customer self-service experiences, simply by describing in plain English what they want the bot to accomplish.”

Amazon are targeting companies that use AWS as their server provider, which is a LOT of businesses. Want to get started? Amazon posted a presentation on that too. Very helpful!

15 free courses to learn AI in 2023

We saw that most clicked recently was a post on 10 of the best tools to up your productivity using ai. It seems there is still a big appetite out there to learn and experiment more.

GenAI company founder Steve Nouri, posted this week on Linked In a comprehensive guide to where to find all the best free ai courses out there. If you like self service & free then this is the list for you.

Courses come from OpenAI, Google, Microsoft, Harvard, Stanford & so on so should be plent of quality there.

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:

Q* - ai moves into a new realm

Reuters this week published what they believe was behind the Game of Thrones style manouvres at OpenAI with Sam Altman’s sacking and then reinstatment back at the helm just days later. The answer is Q*

Q* is the name for the newest evolution of ai at OpenAI where the intelligence can can work its way through mathematical problems. The GPT version essentially learns by rote. It views trillions of pieces of data (ie the internet) and uses that information to simply predict the next most likely word in a sentence based on the context of the question. Basically an interpretation of what it may have “seen” a viable answer elsewhere (or barring that, just what it thinks might be the next best word which is what leads to “halllucinations” or the ai just making stuff up).

“The new model was able to solve certain mathematical problems, the person said on condition of anonymity”. Even though these problems were said to be “on the level of grade-school students” the difference here is the ai was reasoning its way through the problem, not just remembering seeing the problem in the past and re-posting the number.

This is the stepping stone to artificial general intelligence (AGI). “OpenAI defines AGI as autonomous systems that surpass humans in most economically valuable tasks”.

So why did Altman get fired? Researchers wrote to the board to express “concerns over commercializing advances before understanding the consequences”. What are those consequences? “There has long been discussion among computer scientists about the danger posed by highly intelligent machines, for instance if they might decide that the destruction of humanity was in their interest.” 😅 

ODIGEO is doing ai personalisation at scale now

Anyhoo, moving right along, ODIGEO is making fists full of cash with ai 😀 

No Johnny come lately to this ai game, ODIGEO Chief Technology Officer, Carsten Bernhard told the Financial Times conference on the impacts of ai that “for the past ten years, eDreams ODIGEO has comprehensively enhanced the traveller experience by deploying AI across the business”.

The ODIGEO group consisting of brands such aseDreams, GO Voyages, Opodo, Travellink are “powered by state-of-the-art technology that services over 2 billion monthly user searches, transforming the extensive offerings of the global travel market, including countless combinations with 700 airlines, into targeted, personalized proposals crafted for each individual customer:”

Virtual interlining was specifically mentioned as one use case in driving this personalisation. Interlining involves combining different air tickets from unconnected airlines into a single itinerary to get a better route or price - and then of course service the customer when things go wrong. Real time updates of flight services was also called out as part of this.

It should have dawned on most managers in travel by now that ai is the missing piece to unblocking all and any data they may have been collecting and warehousing, either unsure how to utilise it or pouring it like diesel into a petrol powered CRM and wondering why it is spluttering along.

As Bernhard states “There’s no way a human could process this volume of information. We have to find ways to translate that mass of data efficiently into targeted, relevant travel options for consumers. The only way to do this is through AI.”

ODIGEO are already there. How about you?

Travel Counsellors buid their own ai assistant for advisors

In a week where agents at Aito’s Thailand confrence were being urged to “digitise the predictable and humanise the experiential” and an OTA was telling traditional agents “will become “more important than ever“, the UK home agent brand Travel Counsellors were transcending the verbage and kicking into action.

Travel Weekly reported on the iminient release of the TC Co-Pilot which will “produce beautiful results personalised to your customers’ needs”. Chief Technology Officer Jon Bauer was however quick to echo the statements being made over in Bangkok. TC Co-pilot is “not here to replace what you do. We’re a digital business that feels human,” he said. “We’re a personal business in a digital world; we’re a business that augments and nourishes relationships instead of replacing them.”

One of the interesting use case talked about was the ability to highlight products that match the profiles of dormant customers giving the agents a timely and specific reason to reach back out to them.

That sounds like a “£15m investment” well spent.

Travel journalist finds article she never wrote

The curious case this week of a journalist who was surprised to see in her Google alerts on her own (very unusual) name, pop up with an article about a town she’d never been to and never wrote about sitting under her byline on a company’s site she had never worked for.

This is something we unfortunately suspect might become more common place as content teams just “hack” their way to new content to ramp their SEO by using tools like GPT.

As we mentioned above, GPT just predicts what is the most likely next word and when it seemingly came to predicting the name of an author for yet another appalling listicle, it chose real author Iris Dorbian.

In a neat segue, I have my own theory that the decade long blietzgrieg of chasing SEO rankings by trying to out do one another with one appalling listicle after another, is actually what stopped GPT from being better out of the box at creating trip itineraries. As pure lists these articles basically ignore geography and data points like the time it might take to get from place to another, let alone the time you might want to spend there. I suspect listicle reader is only there to take a selfie for their Insta so maybe they don’t care.

Anyhoo, now GPT which is trained on this dumpster fire of “content” thinks this is what we want because it was most prolific on the internet. Bravo! Those training models need to basically beat this mindset out of the ai (and its not easy). We discuss this on my up coming podcast with Shie Gabbai from Roam Around - stay tuned for when that drops.

As for the offending site in this case. They seem to now be having some technical difficulties after Iris contacted their web server provider.

Potentially a good case to test OpenAI’s new commitment to pay any compensation claim made from copyright infringement from those producing content using their service. They snuck this in at the beginning of their Dev Day a few weeks ago to little fanfare but ultimately it could be one of the bigger announcements. Seems like pouring petrol on the dumpster fire to me.

Google ai helps airlines on climate change

Spotted this one this week on airline industry veteran Timothy O'Neil-Dunne’s Linked In. Seems Google has teamed up with American Airlines to find flight paths that reduce the making of contrails.

Admittedly, I wasnt really sure what a contrail was but “contrails tend to form when planes fly through humid areas, and the soot from a plane engine's exhaust crystallizes to form ice at higher altitudes. This leaves behind a streak of white, cloud-like exhaust visible from the ground.” Yep - those I know!

The study has found that ai can find routes that both hugely decrase “these streaks (which) could potentially contribute to global warming by trapping heat that would otherwise be reflected back into the atmosphere” but also not overly increase the overall fuel consumption.

“The results indicate that pilots were able to reduce contrail formation by more than 50%, and used just 2% more fuel, on average, by having to redirect their flight. Only a small percentage of flights would ultimately need to be rerouted, meaning the increase in fuel usage could be as low as 0.3%.”

It wasn’t stated how much contrails specifically contribute to aviation emmissions overall, which seems the critical piece of missing information but at least it is something in an otherwise whole lot of nothing in this area.

What is you ai strategy?

One of my favourite finds of this week was this post from Katalina Mayorga, the founder of El Camino. Katalina we’ve actually featured in one of very first editions if the this newsletter and she continues to genrously share her journey as a regular founder building an emerging travel business.

The beauty of ai is its availability to everyone and its opportunity to flatten the playing field.

In her 3 part series in conjunction with ai founder Kene Anoliefo, Katalina takes us inside her mind and as she works through the process of building out the ai strategy for her company.

If you don’t have a strategy or are not sure where to start, start by following this series and then copy (the process, not the end strategy!!).

As a spoiler, Katalina is starting with her customers and what their specific problems are. Not with what magic tricks the tool is capable of.

Attenborough ai

OK - close your eyes and imagine you are on safari and you are one of the lucky ones to spot a cheetah just lazing on a branch. You’ve got your 200 snaps of exactly the same thing from which you’ll pick 1, or maybe 3 best later. The guide has given you a quick overview. What now?

How about point your phone back at the majestic animal and let David Attenborough take up the story.

This is where the world of multimodal is going. ai vision + speech capabilities bring incredible new layers of UI and UX to the travel experience. Of course we might also just get rid of the phone too and wear a pair of Meta’s Ray Bans or have the Humane pin attached to us like a broach and the whole thing become seemless.

I’m still not sure whether this is just incredibly fantastic or a bit something else, but on balance, great storytelling generally enhances the experience and if people are still revering Attenborough in 2123 through something like this, then that is in the positive column.

There is no reason why a safari company couldn’t be doing this right now. Don’t know where to start? Reply to this email and I’ll help you out. Here is my most reent client review:

“I was introduced to Tony as someone who might be able to assist us with navigating the challenges of a growing travel business, but also to help chart a path for the future. With Tony's vast industry experience and knowledge, I must say that the past month of working together has been absolutely transformative and came just at the right time for our business. If ever someone has under promised and over delivered, it has been Tony in this last few weeks of consultation, going above and beyond expectation.Thank you Tony, I look forward to working together again in future.” - you can see more on Linked In

No review this week here but instead I did see a couple of the better known incumbents in the ai travel planning space resharing this article around how their efforts are defensible.

OpenAI may not be coming for them, but the market is getting crowded especially as most now seem to have at least one eye on the B2B market.

Most clicked last week was my Notion workbook of course! 😅 And then people checking out my Linked In to see what bozo got this all so wrong. The article people were trying to get to was the Linked In post about the Roam Around captial raise.

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