Looking at my travel experiences through a learning designer’s eyes
I’ve been quiet on my socials for quite a while. I decided to give myself some time to rest and reorganise, and I travelled quite a bit too. Now that I’m working on my social media content again, I’ve been looking back at those trips and realising that some of my experiences are relevant to my work in learning design. I thought I’d share a few reflections.
I’ve been using AI for a long time, in different versions and for all sorts of situations. But I find it particularly useful—and interesting—when I’m travelling. It fits the curiosity that comes with exploring somewhere new: noticing things, not quite knowing what they are, and wanting to find out more.
Until recently, I mostly thought of this as part of how I travel. I ask questions, research places and build little guides to share with family and friends. I hadn’t given much thought to how these habits connect to my work as a learning designer and consultant. Looking back now, I can see that they raise familiar questions about how we learn, what helps us understand something and how we access the information we need.
Once I started looking back, those connections became easier to see. There were questions about curiosity, but also about access: what happens when an explanation isn’t available in a language someone understands? There were questions about structure: how do you help people explore without deciding everything for them? And there were mistakes that reminded me how much judgement the person using AI still needs.
This isn’t a review of a particular tool. It’s a personal reflection on AI and informal learning: what I did, what helped, and what those experiences make me consider when designing learning for other people.
Pompeii: a guide when we needed one
One of the first times I saw a real benefit was on a trip to Italy a few years ago. We visited Pompeii, a vast site with plenty of walking, villas to explore, and history and stories around every corner. We had our visit and tickets booked, but the audio-guide queue was long, so I turned to Perplexity.
Within minutes, using a few simple prompts, I generated a guide. It wasn’t just a piece of text: it felt more like a mini app, with a map, places marked on it and stories about the different villas. What I liked was that it included the main points, with enough context to help us make sense of what we were seeing. We could read the explanations or listen to them.
I compared it with other free audio guides and mini apps available for Pompeii. To me, the quality wasn’t far off, especially considering I’d generated it on the spot. That was my impression of how useful it was during the visit, rather than a check of every historical detail.
Looking back, the useful part wasn’t simply that AI had produced a guide quickly. It was that the explanation was available while I had something real to connect it to.
As a learning designer, that makes me think about when we provide information. Sometimes people need an introduction before they start. At other times, the need for an explanation becomes apparent only once they encounter a particular object, decision or problem.
That is the connection I see with support at the moment of need. It doesn’t mean every explanation should be left until the last minute. It makes me ask which information needs to come first, and which might be more useful alongside the experience itself.
There was a practical condition too: internet access. Mobile data and roaming affect whether this kind of support is actually available. A resource that works well at home may be much less useful when you’re standing somewhere with poor connectivity.
AI and informal learning: ongoing travel conversations
For a long time, Perplexity was my travel companion. My actual travel companions are often less familiar with AI, and they’re usually surprised by what we can find out while exploring.
Many AI tools now offer conversational interactions and follow-up questions. My early tool of choice was Perplexity because I liked having linked sources, numbered citations and context carried through the conversation. Those source links gave me somewhere to go beyond the answer, although a citation wasn’t a guarantee that the explanation was correct.
What I liked most was moving beyond a single question and answer. I could ask something, see what came back, and then follow the part that interested me. I didn’t have to know every question at the beginning.
On another trip to Rome, we were walking among the ancient ruins—I think it was around the Roman Forum and Palatine Hill. There was so much to look at, but the official audio guide didn’t cover everything that caught our attention. Sometimes it was just too confusing, and we weren’t sure where we were in relation to the guide. Sometimes it offered just a few sentences about a building; other times, we wanted to know about a different aspect of what we were seeing.
That’s where AI came in. We could ask follow-up questions, look for more stories about a particular building and explore the details that interested us. We weren’t limited to the explanation the audio guide had been set to give us.
That was what I enjoyed most: being able to follow our curiosity in the moment, rather than moving on while we still had questions.
We could also ask about nearby places of interest and connect those questions to our location. That’s how we found Chiostro del Bramante in Rome – a quirky, colourful gallery- along with a few other places we might not otherwise have come across. I don’t really know how the geolocation side worked, but it did work for us: we ended up with a few interesting recommendations. It gave us another way to explore, starting with where we were and what interested us rather than only the places we already knew to look for.
For me, this is a useful example of self-directed inquiry. We were deciding what to investigate, prompted by the place and our own interests. It wasn’t a course, and we weren’t required to cover the same material as everyone else.
In teaching and training, there usually are shared requirements. But that doesn’t have to mean every question is predetermined. I can imagine providing a common starting point, then making room for learners to investigate something they notice, ask a follow-up question or bring a different explanation back to the group.
The part I would want to preserve is the learner’s question. AI could help someone explore it, while a teacher or facilitator helps them evaluate what they find and connect it to the wider topic.
Travelling with my mum: when the explanation needs to be in a different language
On another trip to Rome with my mum, most of the places we visited offered audio guides in foreign languages. I’m Polish and speak English, but my mum doesn’t. When a Polish guide was available, we could both follow the explanations. Where there wasn’t one, the language barrier made things more difficult for her.
Having AI available was amazing in those moments, especially being able to ask questions by speaking and hear the answers read aloud through text-to-speech. It offered another way to access explanations in Polish when the existing guides didn’t meet our needs.
This experience makes the question of access very concrete for me. Information can be available and still be out of reach for the person who wants it.
When I think about this as a learning consultant, I’m not starting with “How much more content could we generate?” I’m asking whether the content we already have is usable by the people it is meant to support.
Language is one part of that. Format and interaction matter too: someone may want to listen, speak a question or read an explanation. Offering alternatives can help, but I wouldn’t describe AI translation or audio as an automatic accessibility solution. The language needs checking, and the format needs to work for the person using it.
For me, the lesson is to begin with the barrier someone is experiencing, rather than with a feature we happen to be able to add.
Bringing our trip research together in one place
I travel quite a lot, usually with family or friends, and I’ve always done some research beforehand. Over the years, that research has taken different forms: printed information, notes, or things I tried to keep in my head. Now, I bring it together in a mini travel guide that I can share with everyone coming along.
It’s not necessarily a fixed plan. It’s a map-based app built around my research and the places that interest us. It includes places we might want to visit, how to get around and useful transport timetables, so we don’t have to search for the same information repeatedly.
If we’re heading somewhere in the morning and returning in the evening, for example, the main transport options and departure times are already there. I also include directions and distances, which help us decide whether we want to walk or take public transport.
Sometimes this becomes a loose itinerary for each day, but I like to keep it flexible. The main points are there to help us make decisions, rather than dictate what we do. We can see what’s nearby, how far apart places are and what travelling between them would involve.
What I find useful is that I haven’t had to change the habit itself: I still research our trips, as I’ve done for years. The difference is how I organise and share what I find. Instead of having information scattered across notes, printouts and browser tabs—or relying on me to remember it—it’s gathered in one app that everyone travelling with me can use. Putting it together doesn’t take much time, and it saves us repeatedly looking things up while we’re away.
These small travel builds were among my early experiments with AI app development. I’ve written more about that in
Lovable for Learning Design: 4 Practical Examples.
This is where I see a connection with designing support around an activity. The guide gives us enough structure to get started and make decisions, without removing the choices that make the trip ours.
I would be careful about calling every useful guide “scaffolding”. In teaching, scaffolding involves support that responds to what a learner needs and changes as they become more capable. My guide isn’t necessarily doing that. But it does prompt a related design question: what support would help someone participate more independently, and what would simply take over their decisions?
It also reminds me to distinguish learning from practical assistance. A timetable helps us catch a train. An explanation helps us understand a place. Both are useful, but they aren’t doing the same job. In a training project, that distinction can help me decide whether people need a learning activity, a reference resource or a clearer process.
Taking a photograph—and sometimes getting the wrong place
Another early example of feeding my curiosity on the spot was taking photographs of interesting buildings and using AI to find out more about them: what a building might be, its history or which details to look at more closely.
What I like about this is that I don’t have to know the name before I can ask a question. If I’m not sure how to describe a building, a photograph gives me a starting point.
But this has also gone wrong. Photographs of buildings can return similar-looking places that are incorrect. A neo-Gothic church or chapel might be matched to something that resembles it but is in a completely different part of the world.
That matters because the answer can still sound convincing. A detailed explanation isn’t much use if it belongs to another building.
The teaching-and-learning connection here isn’t just about making questions easier to ask. It’s about what happens after the answer arrives. Does it fit the location? Do the details match what I can see? Can I check it against a local sign, an official website or another reliable source?
As a learning designer, I would want an AI-supported activity to make room for those checks. Getting a response and evaluating a response are different skills. A learner needs opportunities to notice a mismatch, revise a question and decide when to consult another source or person.
The stakes matter too. A mistaken identification during sightseeing is not the same as an incorrect answer about a workplace safety procedure. The level of checking needs to reflect the consequences, not how confident the answer sounds.
AI and informal learning: reflections for learning design
I don’t think my trips demonstrate that AI improves retention or makes everyone learn more effectively. I haven’t measured that. What they do give me is a set of experiences to reflect on.
The connection between AI and informal learning feels natural: I was asking questions outside a planned course, guided by what interested me and what I encountered. The physical setting gave those questions context. That doesn’t mean standing beside a building is, by itself, a complete example of situated learning, or that receiving an answer completes an experiential-learning cycle. Those theories involve more than a location or a conversation.
For my practice, the useful questions are simpler:
- Where do questions come from? Are we making room for what learners notice, or only for what we planned to explain?
- When does the explanation become useful? What needs to be introduced beforehand, and what could be available during an activity?
- Who can use the support? Are language, format or navigation getting in the way?
- How much structure is enough? Can we help someone get started without deciding every step for them?
- What happens after the answer? Do learners have a way to compare, question, check and apply it?
These are not new questions for learning design. My travel experiences have made them more tangible to me.
They also remind me not to confuse an interesting explanation with developing a skill. Teaching may still need demonstration, practice, feedback and assessment. AI-supported exploration can sit alongside those things without replacing them.
Practical tips for using AI to support travel guides and exploration
Based on these experiences, there are a few things I would keep in mind when using AI to prepare a guide or support questions during a trip:
- Start with the people and the purpose. Our interests, language needs and practical arrangements matter more than which tool happens to be available.
- Leave room for follow-up questions. An initial answer may reveal what we actually want to know. I don’t need to turn every interaction into a finished lesson.
- Keep sources within reach. Where a tool provides links, use them to investigate—not as a badge that makes the answer automatically trustworthy. Where it doesn’t, check independently.
- Check information that affects a decision. Transport times, opening hours, access arrangements and safety information need current, authoritative sources.
- Use photographs as starting points, not proof. A visual match can be wrong even when the accompanying explanation sounds plausible.
- Offer useful formats and check them. Text, audio and another language may help different people, but they need review rather than assumptions about accessibility.
- Plan for connectivity and privacy. Keep necessary practical information available without a live AI conversation, and consider what photographs, voice recordings or personal details you are sharing.
For a learning project, I would add one more question:
what should the person be able to understand or do afterwards? That helps me decide whether exploration is enough or whether the activity needs further guidance, practice and feedback.
Looking back at what I was already doing
I hadn’t set out to test a theory of learning while travelling. I wanted to understand what I was looking at, help my mum access explanations and make the information I had researched easier for everyone to use.
AI was useful because it fitted those needs. Sometimes it helped me explore a question. Sometimes it helped us organise a day. Often it made access to information possible. Sometimes it was wrong and needed checking.
Seeing the connection to my work doesn’t turn those trips into formal learning experiences. It does make me look differently at the support I design: the questions people bring, the barriers they face and the choices they should still be able to make for themselves.
If you teach, design courses or support learning at work, which everyday experiences have made you reconsider how you provide explanations or guidance? I’d be interested to hear the examples—not just the tools.
FAQ
Why did I start with Perplexity?
I liked the linked sources, numbered citations and ability to continue with follow-up questions. It was my early choice, rather than the only tool capable of supporting a conversation. This article is about the experiences and what I take from them, not a recommendation to use one product.
Does asking AI questions count as learning?
It can be part of learning, but receiving an answer doesn’t tell me what I have understood, retained or can apply. In these examples, AI supported my inquiry; I’m not claiming measured learning gains.
Can I rely on AI to identify a building or provide travel details?
I treat a suggested identification as something to check. I’ve received answers for similar-looking buildings in the wrong location. I also check information such as transport times and opening hours against current official sources.
What would I take from this into teaching or training?
I would consider where learners need explanations, how they can ask follow-up questions, whether the support is usable in their language and preferred format, and how they will evaluate an answer. I would still design practice, feedback and assessment where the purpose requires them.
Resources (for more serious reading on the topic)
