Bruce Lee, baby lions & ASK: AI superpowers for attractions

Posted on November 21, 2025

👤 Featuring Ravi Chandra, Chief Product and Technology Officer & Angie Judge, CEO of Dexibit

Angie Judge and Ravi Chandra dive into the explosive adoption of Ask, Dexibit’s AI engine built for visitor attractions. They explore why vertical AI delivers deeper insights than general tools, how Ask retrieves and analyzes complex visitor data, and what AI means for forecasting, simulation, and smarter decisions across the visitor experience. A fast, fun look at the future of AI in visitor attractions.

Transcript (generated with AI)

If you want go from gut feel to insight inspired, this is the Data Diaries with your hosts from Dexibit, Ravi Chandra and Angie Judge. The best podcast for visitor attraction leaders passionate about data and AI. This episode is brought to you by Dexibit. We provide data analytics and AI software specifically for visitor attractions so you can reduce total time to insight and total cost of ownership while democratizing data and improving your team’s agility. Here comes the show!

Angie: Ravi, Justin just sent me our bill for one of our AI providers. Looks like it’s getting bigger and bigger every day.

Ravi: Yeah, it’s looking a bit scary, and I find it funny because right now, globally, we’re supposedly in the midst of an AI bubble that’s about to burst. But what we’re seeing is that we’ve only just launched our new ASK AI-based feature, and already the usage is through the roof. So much so that our AI spend is approaching what we spend on the cloud. It shows two things: one, the power of AI is real, and two, maybe there isn’t even a bubble at all. Maybe this is the future.

Angie: Yeah, well, the usage reports don’t lie. Neither do the bills from the AI providers. And I must admit I’m guilty of this myself—I now just refer to ASK to do things in our data that I would have previously done the hard way. Like this morning, extending the US federal government shutdown in the almanac—very easy to do on the prompt line. Who can be bothered clicking away in a calendar these days?

Ravi: Exactly. It reminds me of something my late grandfather used to tell me. He had this whole Rolodex of proverbs, and one of them was, “Penny wise, pound foolish.” The point is, it costs money, but it’s so valuable that it’s worth it on balance.

Angie: I know where you get your great proverbs from. Was it you who told me this one? Or maybe it’s just something I’ve heard a lot in the tech industry about AI—that people usually overestimate the pace at which they’ll see change, but underestimate how far it will take them. That’s probably true for AI as well.

Ravi: Not from me, but I love it, and I agree.

Angie: I was talking to one of our team as we’ve released ASK. For anyone tuning in who doesn’t know Dexibit well, we do data and AI specifically for visitor attractions. We’ve just released this new part of our product called ASK. It’s conversational insight—a GPT-like experience connected to your data. Our data, your world, your visitor attractions—trained on your voice of the visitor—and with access to the internet and a pile of other cool stuff so you can work with your data in powerful new ways.

We’ve been collecting feedback from our team and customers, and one thing really stuck with me. They said: when you go through dashboards and visualizations, you eventually get to insight, and eventually, after working through that, you get to something actionable. But with ASK, you get to the insight first. Or you get to the action first, and can back out from that to explore the different angles. It speeds up the thinking process. It’s almost like getting the answers before you have the question.

And I think we’re coming across all of these lovely discoveries as we go because the timing is still early.

Ravi: Yeah, I think we’ve found this beautiful window to launch ASK—between “this is possible” and “everyone has this.” It’s an amazing place to be. We feel like we can deliver a really confident, accurate product, and bring it to market at a time when almost none of our competitors have anything remotely like ASK.

In 2023, customers asked, “Can AI do this?” In 2024, it was “Should AI do this?” And now we’re at the point of “Why isn’t AI doing this for me already?” We’re meeting the market at this inflection point.

And I think three things have converged. First, the technology is finally ready. We couldn’t have delivered this quality even six months ago. Dexibit’s first AI in production was January 2024—almost two years of experimenting and learning. Second, there’s general AI fluency—ChatGPT has revolutionized personal computing and replaced Google for many people. Third, we’ve matured our own product—domain modeling, our semantic layer—all the foundations that make ASK possible.

Angie: Exactly. People used to ask me, “Why not supermarkets? Why haven’t you gone after retail?” My answer has always been: you can be more powerful by solving a problem deeply within a vertical. It’s better to dominate your niche than be meaningless to the masses. And in the age of AI, vertical focus is even more important. It allows you to tune so carefully. Everyone can put questions into ChatGPT, but that doesn’t mean you get the right answers.

Data access is one thing, but knowledge is another. Metadata, context, understanding every system and attribute—feeding that in for all our customers with a unified view of the world—that’s incredibly powerful. Comparing that to training an AI purely on the general internet produces entirely different results.

Ravi: Right. Last week we were asked what makes us different from ChatGPT or Claude. It seems obvious to me as a techie, but not everyone sees behind the scenes. ASK is a special‑purpose AI agent that lives inside Dexibit. Dexibit provides the core capability—almanac, multi-location management, user controls, sharing, scheduling, email reports, PDFs—and ASK becomes this super smart artificial colleague that understands every aspect of that world.

We might share underlying LLM technology, but the specialization makes it incomparable. I think of that Bruce Lee quote: “I fear not the man who has practiced 10,000 kicks, but the man who has practiced one kick 10,000 times.” That’s what we do—focus deeply—and ASK is the result.

Angie: Summing up our business model with a Bruce Lee quote—that’s impressive. But yes, the answer isn’t “buy Claude or buy Dexibit.” It’s probably “use both.” If you’re writing a standard employment agreement, sure—Claude or ChatGPT can help. But if you need to know how many visitors you’re forecast to have next quarter, whether you’ll hit your targets, whether you need to reproject, or what your simulations look like—that’s a totally different problem. That’s where ASK comes in.

Ravi: Exactly. Your data has to be collected, up to date, reliable, modeled correctly. Security, governance, permissions. Dexibit already solves those problems—and ASK builds on that.

Angie: And it’s interesting—platforms like ChatGPT are going after PowerPoints and Word docs. That’s great, but it’s entirely different from the quantitative data and voice-of-the-visitor feedback we work with. Different inputs, different jobs to be done.

Ravi: Exactly. People often underestimate the retrieval problem. Collecting data is the easy part. Retrieving what’s relevant is hard. If you’ve got a stack of PowerPoints and spreadsheets, how do you zero in on what matters? LLMs can’t handle huge volumes of content—you can’t throw terabytes of data at them. So how does an LLM figure out what’s useful? That’s the heart of RAG.

RAG—Retrieval Augmented Generation—is really just about finding the right context. A typical LLM can take maybe 100 pages of context—that’s nothing if you have thousands of documents. We use a range of techniques to preprocess and analyze everything—qualitative and quantitative.

Angie: And I know you won’t go into too much technical detail!

Ravi: Don’t worry. But we’ve leaned on established techniques, huge experimentation, fine‑tuning, evaluation. That’s where the real work is. And it’s why some people try Claude or ChatGPT with their own data and quickly discover the limitations.

Angie: Yes—we’ve had multiple prospective customers say they tried uploading a spreadsheet into ChatGPT and it just couldn’t do what they needed. Without the pre‑work, data foundation, onboarding, modeling—you just don’t get useful results. Whereas some of the things customers have done with ASK are incredible—market segmentation or evaluation that would have been impossible for a human.

Ravi: And ASK will continue to get better as we add capabilities and learn the strengths and weaknesses of the models.

Software hasn’t had an off‑the‑shelf pattern for this. When we started ASK, one library we depended on was at version 2. Now it’s at version 6, in just months. Everyone is still figuring this out. For us, it’s been fun—we get to contribute to open source, discover bugs, help build the state of the art. It reinforces that we’re ahead of the curve.

Angie: And the user interface—you’ve introduced such intuitive elements. Being able to call up the Almanac, see item types, view the web search sources—it makes the experience so much more interactive and understandable. And it doesn’t give superficial answers—it gives deep analysis.

Ravi and the team have named our AI “ASK Rex”—Rex the Analyst—after the dinosaurs in our offices and as a tribute to our museum customers. All these little touches make it perfect for the job.

Ravi: Exactly. ASK isn’t just a chatbot—it can control almost every aspect of the product. If your location is changing opening hours, you don’t need to dig through menus. Just ask ASK. We wanted ASK not only to answer but to act.

It eliminates the last‑mile friction. Populate a calendar, create users, schedule reports—it gets the job done without the user hunting for buttons.

Angie: We always said: the more energy you spend getting to insight, the less you have to act on it. ASK accelerates everything. And it enables new kinds of analysis that were never worth the effort before.

My favourite party trick—and Ravi hates when I do party tricks with ASK—is with zoos or aquariums. One job they have is tracking animal births and other husbandry events. Those matter for press releases, marketing, and visitation. In seconds, ASK can create a calendar of all animal births, categories, parents, baby names, naming stories—all info that used to be tedious admin.

Ravi: Exactly. And I’m excited for the future—not just within Dexibit. With our API access, we could eventually have ASK control other applications in an attraction’s ecosystem. One prompt could trigger actions across systems. That would be magical. One of my dreams for 2026.

Angie: Me too. If we can do it in Dexibit, we can do it elsewhere. That’s where the industry is going. Just as visitors will book tickets from a prompt, attractions should be able to run operations from one too.

It also ties into the classic build‑versus‑buy argument. You can spend years and millions building a data warehouse, ML models, benchmarks, visualizations, definitions—or you can have Dexibit turnkey. The same is true for AI. If you build it yourself, by the time you’re done, the world has moved on.

Ravi: Exactly. Dexibit focuses on all the small problems so our customers don’t have to.

Angie: One of the other things that stood out for me in that is how often our customers talk not just about the product, but about the “wrapper” of everything else that comes with it—content, concierge, data success, that knowledge of how you make the most of it, how you get value from it, what you do with it.

And I think that’s super, super important in this age of AI. We always had this problem with data that, in terms of data literacy, data culture, data confidence in a team, you get two types of people: people who are very familiar and comfortable and know what they want, and people who aren’t and don’t.

The same is true when you present everybody with a prompt line, right? It’s a blank piece of paper. They need to know what questions they want to ASK. We’ve seen that already in our user base: you get some people who have all the questions and can just go hell for leather and ask away. And then you’ve got people who aren’t sure where to start.

So by, for example, being able to provide them with some base prompts to work with, showing them how to create a prompt and do so effectively, having suggestions, and being able to suggest as you go—those sorts of things can really help them get there.

Ravi: Yeah, exactly. This really reminds me of the Trolls movie. For those of you listening, I have three daughters, so I’ve seen my fair share of kids’ movies.

Trolls were those little figurines from back in the eighties or nineties, and they’ve turned them into a series of movies. At one point the lead character, Poppy, has this great line I love: “No trolls left behind.” What she means is that although some trolls are really grumpy and upset, she refuses to give up on them.

I feel like we bring that same mindset to our customers here at Dexibit. We understand that AI is this new technological era, but we’re not going to give up on any of our customers. It’s our job to help close that gap with onboarding and AI literacy so they can extract the most from these tools and get real value.

I think that’s going to look like a series of different steps—both innovating in our user interface, making it easier and more accessible for those who aren’t quite sure what to do, and also more general content. I’m probably out of my depth here, but things like webinars, tutorials, and of course our fantastic customer success that we’re known for.

Angie: One of our most popular resources at the moment—this is on our website, resources.dexibit.com if anybody needs a plug—is a template for developing an AI policy. We don’t necessarily recommend that you have to have one, because if you’ve got a great data policy, good security policy, and training and culture among your team, you’re probably safe to fly away in the land of innovation.

But if you feel like you have to, or somebody’s saying you’ve got to—like your board or legal advisor—then it’s there to fast-track that process, because everybody doesn’t need to reinvent the wheel.

You can either spend nine months navel-gazing, wondering what might go on that policy, or you can just do—because that’s really the only way to tackle AI: get out there, experiment, start using it, and see what happens. If that’s something we can fast-forward for all of visitor attractions, then that’s something we will do.

Ravi: Yeah, absolutely.

Angie: What else is on your mind when you think about the future of AI and visitor attractions, Ravi?

Ravi: What is on my mind? There are a lot of things I think about in terms of the future. Maybe this is a great point to drop in one of my favourite quotes ever, from Alan Kay, a computer scientist, who said, “The best way to predict the future is to invent it.” I think this is so true right now.

More concretely, there’s one aspect I’d love us to get into, which is really leveraging ASK and Rex with forecasts and dynamic pricing—bringing some of that deep data science capability into the product and making it really accessible.

I’ll be the first to put my hand up: I don’t have a background in data science or machine learning. But if ASK could help me, I’d definitely use it. I think this is going to open up so many opportunities, especially in conjunction with things like scheduled conversations, for a really pragmatic take on what visitor attractions can start doing with this technology—or start seeing it do.

Secondly, I’d love for us to tailor the experience for every single decision-maker within an organization. Right now it’s quite general-purpose. But with a bit of smarts, we can hone that based on role or persona so people get even more valuable, targeted results, and can take action within a well-understood scope and remit.

I could keep going, but honestly, I should probably stop, otherwise this podcast will be over an hour long…

Angie: And never released, because it will contain all of our dreams and secrets.

But yeah, I totally agree. They almost converge on each other really—from a data science perspective in terms of personalization and prediction, in different ways.

I’ve already seen a bit of what ASK can do when it has access to machine learning forecasts and to goals set by the customer, usually by their finance team. Being able to simulate what happens, together with the voice of the visitor—what happens if we close this, what happens if we do that—and play in that in-between space of simulation for finance teams, marketing teams, and others doing scenario planning as part of their strategic planning… that’s super powerful.

It’s taking us into a space we’ve always wanted to be in for forecasts. Being able to work out all these jobs to be done and use cases, and to help users explore them, in addition to having the flexibility of asking anything at the prompt line—that’s such a cool place to be.

Ravi: Yeah, I know what you mean. I think it’s magical when you see ASK rely not just on one type of data, but on a whole range—visitation, ticketing—then juxtapose that with visitor feedback and organisational context.

So it might know that a certain area of the location was under construction, and adjust its sentiment analysis based on that insight. That’s astounding, because it would take a lot of human effort to do that before, and more traditional purely machine-based methods would probably ignore that nuance.

That’s the real magic with ASK: it can deal with nuance. I think that’s what will really surprise you if you’ve never used it before.

Angie: Mm. I love the way it loves paradox. It will say things like, “The people who give you the harshest feedback are often your greatest fans.” It finds these weird relationships in your data that are really hard to see with the human eye.

Ravi: Yeah, it’s that tough love, right?

Angie: I know you love a good saying, Ravi, so I had to bring one along for the podcast—for anybody who hasn’t put their toe in the water yet with AI and just needs to jump in the deep end. It’s that famous line: “And the day came when the risk to remain tight in a bud was more painful than the risk it took to blossom.”

I think that’s where we are with AI. It’s time to jump in.

Ravi: Yeah. Absolutely. I love it.

Angie: Well, on that note, I’m off to ASK some more questions.

If your goal is to get more visitors through the door, engaging and spending more, leaving happy and loyally returning – check out Dexibit’s data analytics and AI software at dexibit.com. We work with visitor attractions, cultural and commercial, integrating with over a hundred industry source systems across visitor experience and venue operations, providing dashboards, reports, insights, forecasts, data management and a unique data concierge.

Until next time, this is Dexibit!