Since Garmin launched Garmin Connect+ back in March, the reception hasn’t been great – at least on the Internet.
The premium subscription service includes several exclusive features, and the headline feature is “Active Intelligence” – Garmin’s new take on AI. As it is implemented today, it basically summarizes data that’s already readily available in Garmin Connect, and it offers no serious actionable insights.
I signed up for the free trial, and I did not have high expectations. Other than the live tracking improvements (a nice, but minor, quality of life feature), I found no value in the subscription whatsoever – and I’m comfortable saying it’s a complete waste at the moment.
Garmin’s AI implementation was lazy and half baked. But it doesn’t have to be this way.
Although I cringe a bit every time I see someone post about how they’ve used ChatGPT to create a training plan and track their training, it is possible to do useful things with current AI technology to provide value to recreational athletes and help them achieve their goals. And if Garmin (or Strava) were smart about it, they could absolutely dominate that field.
So today I wanted to take a few minutes to think about what AI – specifically current generation LLMs – are good at and what kind of useful applications they might have.
But first, let’s talk about some other Garmin Connect+ headlines that you might have seen in the last few weeks.
Latest Headlines About Garmin Connect+
A couple weeks ago, I noticed some articles circulating on Reddit about Garmin Connect+. It doesn’t take much to get people riled up on this topic, and the general vibe was that Garmin was doubling down on its evil plan to move more features – and people – behind a paywall.
Here’s one headline: Garmin quietly confirms our worst fears about Garmin Connect+, says more features will ‘likely’ be paywalled in the future.
Although most of the article is fairly benign, it includes this zinger at the end: Garmin users’ worst fears, that a growing subscription services will eventually see more and more features hidden behind a paywall, are quickly being realized.
The headline and that conclusion lean heavily into the hyperbole that’s already bubbling up in the online Garmin community. But … it’s non-sensical.
All of these headlines stem from a few comments made by Garmin CEO Cliff Pemble, during Garmin’s Q1 earnings call. One of the callers asked about Garmin Connect+, and here’s Pemble’s full response (emphasis mine):
Yes. I think we’ve been saying for a while that, that we are evaluating opportunities to have a premium offering on Garmin Connect. And I think the developments of AI and particularly around AI-based insights for our users was one of those things that we felt was important to recognize the value for the investment that it takes to do.
And so we felt like it was the right time, and we do have a very strong user base. Connect+ it certainly doesn’t — isn’t required for users, and we’re not taking any features away from people that they’ve had and we still have a strong commitment to develop Garmin Connect and our devices with broad features that are available to everyone.
The certain ones, we will likely reserve for premium offerings and so far, the response has been positive. We’re not measuring success in terms of the short-term. This is a long-term thing for us, a very important part of our Fitness segment going forward. So we’re going to build on where we started.
In that statement, Pemble clearly stated:
- Existing features will not be put behind a paywall
- Garmin remains committed to Garmin Connect as a free service
- Certain new features will be reserved for a premium subscription
To me, that sounds perfectly logical and reasonable. This is not earth shattering, and it doesn’t contradict anything Garmin said when they released Garmin Connect+.
I don’t know where anyone could get the implication from their original announcement that Garmin didn’t plan on introducing new features. That just doesn’t make any sense.
Obviously, if they’re going to offer a premium subscription, they’re going to develop features that they hope will make the subscription worthwhile for users.
I think it’s crazy to equate this with users’ “worst fears.”
Our worst fears are that Garmin puts every new feature behind a paywall and makes Garmin Connect all but unusable without paying a subscription fee.
There’s absolutely no indication that Garmin is moving in that direction, and the release of the Forerunner 570 / 970 is evidence of that. These new watches introduce new features to the Garmin ecosystem that didn’t exist before – and they’re not behind a paywall.
Personally, I don’t ever plan on paying for a Garmin Connect+ subscription – unless they really up their game. But I also think people are blowing it way out of proportion and worrying about nothing.
What Is Current Generation AI Good For?
With that out of the way, let’s shift gears. What is AI actually good for – and how could Garmin make it useful if they really wanted to leverage it to make Garmin Connect+ a worthwhile product?
One thing LLMs are good at is summarizing data and converting that to text. That’s basically what Active Intelligence does now. It takes some seemingly random metrics that exist in Garmin Connect and plays a kind of Mad Libs to come up with a summary.
The problem is that this adds no value. We already have a really good way to summarize key metrics and data points – Garmin Connect.
Whether you look at an activity in the app or flip through the data screens on your watch, Garmin has already boiled things down to a few key metrics. These are displayed visually, and it’s easy to grasp what’s going on. Personally, I found those textual summaries harder to make sense of.
Instead, if they leveraged these AI capabilities, they could come up with something useful, including:
- Access larger datasets and identify useful insights
- Offer people choices and help them take action
- Answer questions and engage in conversation
- Allow people to interact with the system through natural language
What would that look like in the context of Garmin Connect?
Make Sense of Large Amounts of Data
Telling me the pace or cadence from a recent run is pretty basic and useless. I can already see that easily in the app.
AI is great at taking in large amounts of data, combining it, and producing a summary. Instead of telling me what my pace was on today’s run, how about telling me how this run compared to similar ones?
If I’ve run a particular route frequently, I’d like to be able to make comparisons. But in addition to just showing what my time was in previous runs – which is pretty basic – you could identify factors that may have influenced the outcome. Was the weather different? Was my sleep worse this week than previous weeks?
If I was slower today compared to other similar runs … why was that?
The great thing about the Garmin ecosystem is that it contains a lot of rich information about your health and training. Your activities, your heart rate variability, your sleep, the weather, your acclimation to that weather.
For another example, I recently ran a marathon (Jersey City). In the weeks after the race, I was curious how my recovery compared to previous races. I had to manually scroll back through multiple graphs to see my HRV, resting heart rate, and Body Battery, along with my activities, to periods that I happened to know followed a marathon.
With the data available in Garmin Connect, it would be trivial to identify a time period (two to four weeks) after a specific type of event (a marathon) and compare some of the stats that Garmin tracks. That’s the kind of insights that would be useful – not telling me that my HRV is lower today than it was a few days ago.
Offer Choices and Move People to Action
It seems like forever ago, but Netflix became popular by solving a simple but complex problem – how to recommend things that people wanted.
When I was younger, if you wanted to watch something on TV, you’d either a) flip through a limited set of channels or b) flip through the TV guide to see what was on. Now, there’s so much content that it’s hard to know where to start … unless someone gives you options.
Whether you’re looking for a workout to do or a new course to run, AI would be useful to take a plethora of options and narrow it down to three choices.
Right now, there’s an auto-route generator in Garmin Connect, but it’s of limited use. When I’m out of town, I usually look around Google Maps to find good parks or places to run. I’d love to be able to open Garmin Connect, say I want to go for a run, enter a few details, and get three nearby routes to choose from.
Similarly, if I’m working through a coaching plan, maybe I’m not thrilled with the exact workout for the day. Create some variations on that workout and let me choose from one of them.
Someone who understands running can do that easily. Jack Daniels’ Running Formula is basically a cookbook for workouts. But this would be really useful for the average athlete who isn’t that versed in workout construction.
Answer Questions and Engage in Conversation
To me, this is the biggest oversight.
People love chatbots because … you can chat with them. You can ask ChatGPT a simple question and get an answer. Or you can continue that conversation.
On a basic level, offering users some pre-generated questions that they can ask about their data would be useful. Instead of generating a single summary – and hoping you identify what people want to know – offer them up some questions.
In my day job, I work with Tableau. And this is one of the basic AI-powered functionalities of Tableau Pulse. You create a dataset, identify some important dimensions, and users can use pre-generated questions to explore the data. It isn’t earth shattering, but it can be more useful than letting someone run free with the data and without guidance.
On a more advanced level, people could just chat with a Garmin Connect chatbot that had access to their own data. Ask the questions you want. Get feedback. Get support.
Personal coaching is labor intensive, and that means it’s expensive. If you’re a runner and you want a coach who will pay any serious amount of attention to you, it’s going to cost big bucks. And if you can afford to spend $100 (or a few hundred) each month, great!
But if you can’t, you don’t have an option (other than chatting with ChatGPT). This seems like such a natural extension of the Garmin Connect platform that it should absolutely have been included in the first iteration of Garmin Connect+.
Allow People to Interact with Natural Language
I saved this one for last because it relates to all of the previous use cases.
Any of the things I mentioned before could be done without AI – or without really leveraging an LLM. AI might help the workflow so that you have to pre-program less stuff. But it’s all possible.
I’m a big fan of the Decoder Podcast with Nilay Patel from the Verge. And something he repeatedly brings up is the idea that the core innovation of current LLMs is that they create a new interface. It can totally change the way we interact with computers – or tech in general – because we can interact with natural language.
All of the buzz today in the AI world is about agents and “agentic AI.” You create an agent and send it off to do things for you.
But the idea of having a computer do something for you isn’t new. From the dawn of computing, it’s always been about automating tasks that you’d otherwise have to do yourself. The way we’ve interacted with them has gotten simpler and more intuitive – from punch cards to a terminal to a graphical user interface – and letting people talk to their computers in a conversational way is the next big leap.
Garmin Connect already has training plans, performance metrics, and routes. It collects data from you about how your run was, and it knows what the weather will be like tomorrow.
It already does, to some extent, everything I listed above. What it doesn’t do is let you chat or talk with it like a human coach.
Will Garmin Connect+ Ever Be Useful?
Yes, it makes sense that Garmin will create new features and put them behind a paywall. They’re looking for simple wins, and they want to encourage people to subscribe. I get that.
But new features are just little add-ons. They make things incrementally better. And adding new features behind a paywall pisses off (some) existing users.
The thing that would make Garmin Connect+ truly useful – and create a qualitatively different experience – is to use AI to be a natural language bridge between the user and the platform.
Once Garmin – or somebody else – essentially creates an AI coach that can synthesize data, talk to you about your progress, answer your questions, and manage your training, they’ll be on to something. Until then, they’re just spinning their wheels and wasting money.