WhatDo is a new travel planning tool leveraging AI to curate personalized travel itineraries in a range of global cities to recommend unique experiences, attractions, and dining spots specific to individuals' interests.
Plan and book urban adventures and everyday experiences such as planning trips to the beach, hiking destinations, and culinary exploration suggestions in food hotspots.
+ 352 others joined this month
Comments
Hey everyone!
I'm a young web and mobile developer from Croatia, passionate about creating intuitive and functional digital experiences. I'm self-taught, driven, and always on the lookout for opportunities to grow my skills.
WhatDo uses data-driven recommendations and insights from travel experts. The goal is to create not only a personalized but also a well-informed travel plan through a seamless and efficient planning method.
The platform was created from a team of tech innovators and travel enthusiasts seeking a streamlined travel planning solution that invigorates rather than drowns the excitement of travel.
👉 Try it out today at whatdo.in
Looking forward to your questions and feedback!
Hi Filip,
Thanks for sharing WhatDo with the Travel Massive community (and welcome!)
I have to hand it to you, I wasn't going to publish yet-another-AI-trip-planner — however you've nailed the design so I think it's worth sharing as what you've created is very well polished.
A few questions.
1. Can you explain a bit more about how you use "recommendations and insights from travel experts" to build the product? Is this just a bit of jargon for having used LLMs that have scraped travel blogs (e.g. OpenAI) or are you doing something specific with human input? Please explain.
2. Partners. Are you only linking through to Viator or do you have additional products / experiences sources from other sources? e.g. If you're only recommending Viator products then is this just a "Viator in disguise"? Do you have plans to add additional items such as hotels?
3. How long did this take you and your team to build? What features are you looking to build next?
Thanks again for sharing, and I hope it's a success! I really like the "Take me anywhere" link, too.
Best,
Ian
Hi Ian,
Thanks for the kind words first of all. Of course I can answer all of your questions :)
1. You kind of got it right with the jargon thing, but we actually try to reach out to as many travel experts/guides etc and ask them how they view our website and what feature would they like to see and what is missing. We are utilizing AI of course since at the end its an AI travel planer but once we get the feedback for a certain city/country/day plan we easily adjust it to be more accurate and help as much people as possible. WhatDo does not use any kind of scrappers its just smart use of AI for some sections and something is human written. We currently have over 50k cities so it does get a bit hard to manage all of them and 100% personalize all of them. Btw very soon we will most likely have all 4M cities in the world.
2. We are currently partners with Viator, Expedia, Vrbo and Unsplash. For every city and country if you scroll a bit near the bottom you will find "Hotels and Apartments" section. It used to be at the top of the page but we had to "Hide" it a bit near the bottom to decrease the bounce rate for now since we are testing some metrics.
3. Its a team of 3 people behind this, and it took us about 4-5 months of actual work since we haven't worked on this full-time because we all have our 9-5 full time jobs.
Features we plan to add ->
- We just added "Blogs" 2 days ago and so far they are performing super well.
- Threads for each city/country where travelers around the world will be able to chat/comment/like/share everything about that city and maybe get some more insights from locals. For example this will be on /[city]/t route. Our plan is to release it relatively soon as we already started working on it.
- Fully customized chat ONLY for travlers which would most likely be integrated with Viator/Expedia/Vrbo/Booking etc.
- After those 2 features are about we planned to ask the community and our existing users what they would like to see next.
Yea "Take me Anywhere" is a new "feature" which is super simple but makes it so user can discover unlimited amount of places. We sadly limited it to just countries for now.
One more time thanks Ian for sharing this with your community, we will monitor is closely and if you guys want anything new added don't be afraid to ping us or send us an email at hello@9.agency or filip.halapir1@gmail.com whichever you find better works for you.
Hi Filip - the site looks really good but I can't seem to scroll down the page - although that might be a fault at this end and not yours Cheers Stella
Hi, thanks for the upvote and kind words. We have never heard or seen that issue/bug. This is most likely the issue on your end :)
Well, this is one of the first travel products that I've come across that boasts "AI" involvement that doesn't _feel_ like it's fresh off the `completions` API, or, worse, make users suffer through some sort of prompt. Leaving AI out of the picture altogether, it's actually quite good!
I've had to check out the Dublin references of course. While entirely a Dublin dish, "Coddle" is more home cooking and rarely seen on menus. I consider it at the "haggis" end of the scale personally! Two of the "unique experiences" involve getting out of the city completely (to Belfast or the Giant's Causeway), which isn't actually a bad idea for a first-time visitor to Ireland, but perhaps overlooks the option to find more graceful parts of the capital to enjoy. You do score points for suggesting the Bunsen and Fumbally spots.
Likewise for Galway, known as a cultural hub with plenty to do, has experiences hours away, and wouldn't necessarily be my first choice of day-trip either.
I'm guessing that the revenue model is based on selling the tours and activities content that's listed. It's a decent and slick integration, but I think it might be influencing the above results.
From a data science perspective, it would be useful to make sure you're capturing data that can re-train the output to be more refined. This might mean allowing users to share a preference or rate the suggestions (there are many ways to achieve this).
Overall, certainly 5 stars for a v1 and I hope it gets plenty of traction!
Thanks for the detailed feedback Kevin.
Suggestions make sense and it’s definitely something we will be experimenting with.
Cheers, Filip