EPtalk by Dr. Jayne 8/17/23

Many companies are talking about the work they are trying to do using artificial intelligence, but some of the writeups I see range from overly vague to downright nonsensical.

I was intrigued to see an email from Tripadvisor inviting me to “kick-start your planning” by trying the company’s new AI trip planning product in a pre-launch version. It promises to use the answers to four questions about a proposed trip to “sift through over one billion traveler reviews and opinions.” From there, the user is supposed to get a custom daily itinerary based on their interests, and it’s supposed to be sharable with others. I decided to check it out and see how a more consumer-facing offering behaves compared to the healthcare-centric products I usually look at.

I ran the AI trip builder against a trip I had recently planned to a city where I’ve wanted to go, for which I have already compiled a list of recommendations. The solution asked me where I wanted to go (city and state) as well as proposed dates and whether I was going solo, with a partner, with friends, or with family. From there, I was asked to choose attractions, hidden gems, museums, history, culture, food, outdoors, wine and beer, arts and theater, or adventure and sports as the ways I would want to spend my time.

After about a minute, it put together an itinerary, tagged as “powered by AI,” for five full days of sightseeing. I was pleased to see that two of the restaurants on my list matched the Tripadvisor list. However, there was a lot of detail lacking. For example, there was no input for how I preferred to get around – would I have a car, or rely on public transportation? Based on the distances between recommended activities, I suspect it made the assumption that most travelers in the US have a car.

The way that different locations was grouped together didn’t make a lot of sense. I would have expected that AI should be able to use addresses and directions to put things together in a way that makes more sense. Given the amount of data available to the system, it could also have generated some other useful information, such as a forecasted budget per day and a total for the trip. It didn’t ask any information about cost, and also didn’t include any lodging options, which for me is a big part of planning any trip.

Hopefully more bells and whistles will come in the released version. But overall, it didn’t feel like there was much difference between this and other itineraries I’ve seen in travel magazines that I read while researching my trip.

I ran it again against a trip I took earlier this year and the results were even more erratic. It sent me to the same historic district on three different days, recommending that I visit adjacent properties separately even though they were within a block or two of each other. I double-checked the sites’ operating hours to make sure that wasn’t a factor, but they were all open all of the days. It also included the Cracker Barrel Old Country Store, which isn’t necessarily what I think of when I’m looking for great food on a vacation, although I do always like biscuits and gravy.

Maybe some additional functionality could include an understanding of where the trip is in relation to where you reside, whether you’ve been to the area before, and if so, what you did last time. To truly take it to the next level, it could ask if you want to eliminate options that are also available within 10 miles of your home location.

Alas, I’ve spent enough time in this industry to understand the concept of the Minimum Viable Product, so if they were looking at providing something with features that might draw in an early adopter type user or might help validate a product idea, they’re on their way. Unfortunately, there wasn’t any way to provide feedback, so I think that’s a missed opportunity for the Tripadvisor development team. Hopefully they are doing some additional testing where they’ll actually get user feedback. In my experience, that’s the best way to evolve a product.

From All Work, No Pay: “Re: HIMSS selling its conference. It sounds like there might be financial troubles, since the organization is not making good on honorarium payments for CIOs and others who participated in focus groups at the HIMSS23 conference. Staffers have cited the conference’s transition as a reason for the delay.” These focus groups are usually hosted by HIMSS Corporate Members, who pay a hefty fee in exchange for the opportunity to either pitch to or pick the collective brains of various healthcare executives. Attendees usually receive a gift card in addition to lunch. I’ve been to HIMSS meetings where it’s hard to find lunch without waiting in a huge line, so I can see the draw of making the trek to one of the private meeting rooms, which are usually in the basements of the convention centers. A sandwich and the opportunity to rest your feet for an hour can be a draw on its own, not to mention the possibility of seeing something new that a vendor might be preparing to debut. Hopefully HIMSS will get its act together and make good on its agreements if indeed there’s a problem.

I’m always looking for a feel-good story, but unfortunately I ran across this one a few years too late. The Swedish Number was a campaign by the Swedish Tourist Association that allowed callers to dial Sweden and “get connected to a random Swede and talk about anything.” It was launched in 2016 to celebrate the 250th anniversary of a law to abolish censorship. It generated nearly 200,000 calls in the 79 days it was open. Callers from the USA were responsible for 32% of its volume. Swedes could register as ambassadors, and when calls came in, they were randomly routed to participants until someone answered.

It’s a great idea to generate interest in your country, and I like the idea of people just wanting to connect with someone in another part of the world. With all that we’ve collectively been through over the last three years, maybe it’s time for another feel-good moment. Which country wants to give it a go?

If you could call and talk to a random citizen of a given country, who would you want to talk to? Leave a comment or email me.

Email Dr. Jayne.