A word from our sponsor!

With your website - do you feel the need for speed? Whilst speed is definitely important, Brennen Bliss from Propellic gives this advice: “I’m in total agreement with Henry Smith. People will do anything to squeeze out that last ms of speed when they could be working on more impactful decisions if their site is already satisfying user speed expectations.”

Smith stated: “I've audited hundreds of websites over the last 12 months and I rarely see a correlation between speed metrics and rankings. Instead, links, structure, content, and crawlability are the key things that often move the needle. As long as your site is fast enough that the pages render and load quickly, I wouldn't worry too much.”

Need someone looking out for your online health? Check Propellic out for yourself here.

(A big thanks from me to Propellic who have extended their sponsorship of this newsletter through Q3 - great validation & much appreciation).

Do companies really need to hire content marketing agencies anymore? 

This is the question posed by Peter Caputa, the CEO of Databox. Actually - its an open ended question and we’d like to know the answer as it stand for us here in the travel industry. So we’ve partnered with Databox to see if we can help get a statistically significant sample size to find and share some learnings.

Here is the link. If we get 100 respondents then Databox will break it down for us and I’ll have the results for you all next week.

The findings from other industries Perter summarised as:
After reading through many answers, it's obvious that companies are already saving ⏰ & 💰 by leveraging AI in their content production processes.

The public fundamentally doesn't trust generative AIs

This was the conclusion from a great article on the newsletter written by Thomas Baekdal. Whilst the use case in the newsletter was media, I think there is certainly some of the tarnish coming off the AI hype cycle. Ironically enough, a lot of the trust issues with AI, according to Baekdal is that humans themselves are sometimes just a bit 💩 

Deep fakes, fake news, content farms for click bait were all some of the things that the general public are beginning to associate with AI. “When we look at the many forms of AIs that publishers use, the ones they see the most are the bad examples of AI, where publishers are merely using it to get cheap clicks.”

Over in politics, Open AI this week shut down access for two political candidates doing things they shouldn’t with the tech.

In chats I’m having with some founders, they are now not really talking about the AI in their products to their users (but surely still talking it up with their investors?). I also saw a badge this week you could put on your content to say it 100% isn’t AI. It probably also doesn’t help a massive amount in the perception game when Booking.com is sounding the alarm on AI enabled travel scams.

A few other pieces this week also gave a “cool your jets” feel to the current moment. Business Travel News this week reported we are in the age of weak AI. “Jack Staehler, CWT's chief technology officer for counselor and traveler technologies, said at BTN's recent Tech Talk event in Chicago. Current AI technology—including ChatGPT—is classified as "weak AI," meaning it is capable of a specific task but not capable of truly human-level thinking and learning beyond that task. As such, there is a lot of "FOMO, inflated expectations" and a "trough of disillusionment" around AI promises, he said.” It was interesting timing to make these comments given CWT also released this guide into all things that are great about AI in Corporate Travel.

Over at Travolution they spoke to Cloudbed's CEO Adam Harris & “explored this disconnect, highlighting both the potential and the roadblocks to AI adoption in hospitality.”

“Harris acknowledged the confusion surrounding AI's complexity, hype, and potential risks – and that this stands in the way of adoption for hotels. While some advancements are happening in travel planning (think AI-powered itineraries), the hospitality industry itself hasn't seen a breakout moment yet.”

Everything is fast in the world of AI, including the rush to the trough of disillusionment.

What are guardrails for AI and how do you use them?

Amazon has put together a comprehensive piece of information (& big ad) for their guardrails for Bedrock (their AI platform) under the heading of “Safeguard a generative AI travel agent with prompt engineering and Guardrails for Amazon Bedrock.” If nothing else it sounds like it is right in the wheelhouse for our topic here!

The original is full of examples, graphics and diagrams and if you want to know how to put something together (or at least know what questions to ask of whoever is putting something together for you), it is worth looking through the whole thing.

You can find the TL:DR here:

Travel companies are using generative AI for virtual travel agents to enhance customer experience.

Benefits: Personalized recommendations, efficient query handling, and seamless booking guidance.

Challenges: Risks include harmful/bias outputs, exposure of sensitive information, or malicious use.

Solution

• Prompt Engineering:

  • Craft prompt templates to limit responses to travel-related topics.

  • Example: “You’re a virtual travel agent. Only respond to questions about {topics}.”

• Built-in Guardrails:

  • LLMs have predefined responses to block certain queries (e.g., refusing illegal activity instructions).

• Amazon Bedrock Guardrails:

  • External validation of user inputs and LLM outputs using topic denial, harmful content filters, and sensitive information masking.

  • Customizable and integrates with all LLMs on Amazon Bedrock.

Implementation

• Multi-layered Protection:

  • Prompting Protection: Embeds user input in a template to restrict response scope.

  • LLM Guardrails: Built-in measures within LLMs for basic protection.

  • Bedrock Guardrails: External checks for harmful content, sensitive info, etc.

• Monitoring:

  • Track activation of safeguards via Amazon CloudWatch.

  • Logs interventions to adjust and refine filters.

Guardrail Configuration

• Denied Topics: Block finance and politics.

• Content Filters: Block sex, violence, hate, misconduct, etc.

• Word Filters: Block competitor references and profanity.

• Sensitive Info Filters: Anonymize personal info like age and email.

Monitoring

• Use Amazon CloudWatch for logging and creating metrics to monitor safeguard activations.

• Example: Filter triggers and intervention patterns for proactive adjustments.

Cost Estimate

• Roughly $20 per month for 1,000 interactions using services like Amazon Bedrock and AWS Lambda.

Cleanup

• Delete guardrails, CloudWatch dashboard, and metrics created during setup.

Responsible AI Considerations

• Human oversight, continuous monitoring, and transparency.

• Privacy and data protection compliance.

• Inclusion of diverse perspectives in AI system evaluation.

• Ethical training for teams and collaboration for knowledge sharing.

Yale gathers 200 high powered CEO’s in an AI Summit

When you are Yale you can gather together a pretty amazing assortment of CEO’s for your virtual summit. Everyone Booking.com to Goldman Sachs got a run at the recent forum.

I’ve pulled out some of the highlights from Travel and Tourism as well as a couple of insights that we might also be able to learn from:

### Travel, Tourism, and Hospitality

1. Booking Holdings: Personalized Travel Recommendations

Glenn Fogel, CEO of Booking Holdings, highlighted AI's role in enhancing travel experiences by tailoring recommendations to individual preferences. Fogel noted, "Instead of looking at brochure after brochure, AI can offer an even better match with what our customers want to do. It knows your preferences, it doesn’t forget what your preferences are, and it’s all in the database." This capability significantly improves customer satisfaction by streamlining the search and booking process.

2. Liberty Media: AI-Generated Itineraries

Greg Maffei, CEO of Liberty Media, discussed the integration of AI with TripAdvisor to create personalized travel itineraries. This tool, currently in beta, sifts through over a billion user reviews to suggest activities and hotels based on user interests. Maffei emphasized, "Our new AI travel itinerary generator can tailor experiences based on a customer’s stated interests, using the vast data available from user-generated reviews." Itinerary generator! 😄 

3. American Airlines: Streamlined Reservation Processes

Doug Parker, former CEO of American Airlines, shared how AI is used to optimize flight reservations. However, he cautioned against full automation in certain roles, stating, "Nobody seems comfortable with positions such as mechanics and flight attendants being totally automated away." Yes good call on that one Doug (for now).

4. Winnebago Industries: Digital Integration in Outdoor Experiences

Michael Happe, CEO of Winnebago Industries, highlighted the role of digital transformation in enhancing outdoor travel. He mentioned, "We have developed the ability to connect consumers’ phones to systems on our product, so they can operate them remotely," improving convenience and comfort for RV travelers. Not super life changing? Is that even AI?

### Insights from Other Industries

1. Bank of America: AI in Financial Services

Brian Moynihan, CEO of Bank of America, illustrated how their AI assistant "Erica" enhances customer interactions by managing complex banking queries. Moynihan explained, “We had to build a bespoke banking language. It is not ChatGPT, but it is a core application which is a roadmap for ways we can use AI technology.” This highlights the importance of customized AI solutions tailored to specific industry needs and probably something a lot of readers are working on themselves?

2. General Motors: Autonomous Technology in Vehicles (just for Alex Bainbridge)

Mary Barra, CEO of General Motors, discussed the integration of autonomous technology in EVs, aiming to reduce human error in driving. Barra noted, “The ultimate piece of that is autonomy. 99% of fatalities are caused by human error; autonomous technology will help us eliminate that.” This demonstrates the potential of AI to improve safety and efficiency in sectors beyond travel, offering lessons on integrating AI for enhanced operational safety.

(Long overdue) British travellers to face AI lie detector test when entering EU

We all know those pesky Brits are probably hiding something. Right now it is the morbid fear of yet another Cup failure I suspect, but that well founded nervousness could well be misconstrued as something more sinister as they twitch nervously when filling out the immigration card to enter Mallorca as reported this week by The Standard.

“It is understood the technology will be phased in after an initial Entry-Exit System - due to affect all British travellers from October 6 - is implemented, along with a visa-waiver scheme called the European Travel Information and Authorisation System for short stays of up to 90 days.”

More seriously though “Experts warn the scheme, which the EU is considering, could lead to widespread rejections of visas, including discrimination against those with disabilities.” This is exactly the type of bias potentially held in the core of an AI’s training that we need to be really careful to avoid.

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:

McKinsey survey offers new insights into traveller perceptions of AI

Global Consultancy McKinsey have conducted a traveller survey with “more than 5,000 travelers across geographies and generations….. from China, Germany, the United Arab Emirates, the United Kingdom, and the United States.”

Some highlights were:

• “80 percent said that they would be interested in trying to use AI or gen AI to plan a trip. So, there’s an expectation that the use will grow.”

• Chinese travellers are more likely to head to bucket list places (69%).. “For North American and European travelers, it’s only 20 percent” (which seemed low to me) & “travelers from the United Arab Emirates, there is also a strong preference, closer to the number of Chinese travelers, to visit iconic destinations. But what they’re really after are active, sport-heavy holidays—being outdoors, hiking, and doing some sort of exercise”

• “Seventy percent of our respondents mentioned negative experiences related to overcrowding in their travel in the last 12 months.” Add that to the fact that global traller numbers could double in a short amount of time - and we are in for one hell of a squeeze!

Slack Group member Jeff Kischuk mentioned the standout piece of data was that McKinsey “identified seven clusters of travelers, all of whom share a lot of the same attitudes and motivations toward travel. These include sun-and-beach travelers, culture-and-authenticity seekers, strategic spenders, trend-conscious travelers, cost-conscious travelers, premium travelers, and adventure seekers.” This kicked off a discussion on LinkedIn as to how to identify which bucket your customer might be in at any one time in order to personalise for them.

Of course there is always nuance in these groupings as was discussed by the panel of McKinsey team mates in the podcast you can listen to when they said of the the “cost conscious” for example “They might come on a low-cost flight, and they might stay in cheaper accommodations, but they will spend the $500 entrance fee to go gorilla tracking and have that experience.”

New LIVE show where you can ask your AI questions

We’re trying something new out with some LinkedIn LIVE shows so you can ask your burning questions and get an answer (or will it be THE answer!!). I’ll team up with a range of experts for these with the first one being with Christian Watts (who has also already been a guest on the podcast but a lot has happened since then). Christian is as deep as anyone else I know with all things AI and if no-one else shows up - then I’m sure we’ll still have a fun chat about what we are seeing and learning!

The first one is next Tuesday afternoon Australian Time but you can see the timing in your part of the world here.

Slack Group!

How to work with Tony

The calendar is now very full I’m afraid.

Lots of work going on to launch the marketplace for buyers and suppliers to find each other with ai solutions. If you have an ai solution in market (product, SaaS tool, service) - please get in touch to get information about being listed. We are grandfathering forever the lowest listing price for those who are in for the launch.

Please email me to deep dive into the specific opportunities within your business you might wish to explore. By interviewing key internal stakeholders we can identify which of your bottlenecks are most ripe for an ai powered fix and the approach to take to fix those across a month long project. The earliest I am available for this is now September with very limited availability.

For the fully committed business who now understands the transformative power of this technology, the final phase is to move to build your own internal “AI centre of excellence” which is combination of building an ai culture in your business by taking a human centric approach as well as building out or buying in the best solution to each identified issue. Please email me for more details on any or all of these phases.

Want to follow in Propellic’s footsteps and get in front of a highly engaged audience of travel decision makers by sponsoring the newsletter? We are booking Q4 sponsorships now. Also email me on that one for rates and details.

Most clicked last week was the link to the Soul Machines AI avatar, Vesper. 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