'Location-informed' could be the next frontier for generative AI

Want a better way to get from A to B? Just ask OpenAI’s ChatGPT. Thanks to TomTom’s first-of-its-kind large language model (LLM) plugin, ChatGPT Plus subscribers can now use the AI to plan trips, discover new places and explore their world. We sat down with two of the developers behind this pioneering piece of tech to get the inside scoop on how TomTom is bringing its unique geolocation smarts to the ultra-popular AI chatbot.

By now, many of us have had a go at using ChatGPT to help brainstorm ideas, draft emails and write code. For TomTom’s development team, however, the platform has always held a deeper kind of potential – the power to take geolocation services to new places.

“ChatGPT had been on our radar for some time,” recalls Cezary Draus, Staff Software Engineer at TomTom. “When OpenAI announced that they would be adding support for plugins in March 2023, we jumped at the chance to create our own plugin.”

The dev team knew that TomTom’s map data and location services had the potential to make ChatGPT even more useful for navigation, travel planning and exploration – and they were eager to turn that vision into reality.

Marcin Kmiecik, VP of Software Engineering at TomTom, confirms: “We knew that there was real value in a plugin that brought more location awareness to ChatGPT. We also knew that we probably weren’t the only company working on such a tool. It was a race against time to beat everyone else to the punch.”

The team threw themselves into the development process at a blistering pace. And their hard work paid off. In July 2023, just months after OpenAI’s announcement, TomTom became the first company (according to our research) to release a location-informed plugin for ChatGPT. It allows ChatGPT Plus subscribers to access TomTom’s geolocation capabilities, like geocoding, map visualization and routing, right from within the familiar chat interface.

What does that mean in practical terms?

For one, TomTom’s plugin adds an all-new dimension to ChatGPT’s results, going beyond standard text-based responses to display locations and routes on a map. If you’re planning a road trip from Sydney to the Blue Mountains, for instance, not only can ChatGPT give you turn-by-turn directions, it can also show you the way on a map – making those results instantly more helpful and visually appealing.

The plugin also gives ChatGPT access to TomTom’s rich store of geolocation information, so it can provide more accurate and detailed answers to location-based questions.

This is where the magic of location and generative AI begins to reveal itself. LLMs like ChatGPT shine at sifting through mountains of information to surface relevant and useful insights. And their conversational interface allows them to deliver those results in a very natural, approachable way.

When you bring those powers to complex tasks like trip planning or travel recommendations, it makes the entire process much quicker and easier for the average user. As Marcin Kmiecik puts it: “These are tasks that neither technology – generative AI nor geolocation – can solve on their own. And that’s what makes bringing the two together so powerful. Complex location questions are much better with our plugin. You would need to use many other applications to get to the same answer that we can now provide directly within ChatGPT.”

Things get even more interesting when you start adding other plugins to the mix. For example, using a combination of plugins from TomTom and a real estate company, you could ask ChatGPT to show you a selection of properties for sale in downtown L.A. and plan the optimal itinerary for visiting them all in one day. Similarly, with plugins from TomTom and a food delivery app, ChatGPT could serve up a list of pizza places near your location and show them on a map – making it easier for you to find the perfect place to order lunch.

Cezary Draus believes the benefits will multiply exponentially as more plugins become available, allowing developers to tap into a huge ecosystem of capabilities and helping users to complete ever-more complex, multi-step tasks without having to hop between different apps.

He notes: “Perhaps the biggest value of ChatGPT plugins lies in their potential for integration. It creates a virtuous cycle where we can empower other plugins with ours and vice versa. Users will be able to take advantage of these complementary capabilities in one place.”

With one eye firmly on these future possibilities, the development team is keen to keep building on its early momentum in ChatGPT – with several new projects already in the works. While the team remains tight-lipped about the finer details of their plans, they hinted that they are looking at taking LLM-based location insights beyond their current browser-based form, opening new ways for people to navigate the world around them.

Looking to the road ahead, Marcin Kmiecik concludes: “There are so many open possibilities with generative AI; it really feels like we are on the verge of a new era. I'm very proud that TomTom has boldly stepped into this uncharted territory. Thanks must go to our entire team – Yu Guo, Cezary Draus, Ivan Diaz, Massimiliano Ungheretti, Mick van Hulst and countless others – for helping us make such a momentous move; one I hope will be the first of many more to come.”