The Google personal health coach is the first fitness tool that’s actually helped me enjoy Thanksgiving without completely derailing my progress toward my goals. After five weeks of consistently using the AI-powered coach in the Fitbit app, I feel fitter and more in control than when I started.
Built on Gemini, the personal health coach transforms the standard Fitbit app into an AI-driven experience. It’s currently in public preview on Android, with iOS support expected soon, and a full rollout planned for next year. Access requires a Fitbit Premium subscription ($9.99/month).
Although it pulls data from a synced Fitbit device, you interact with the coach exclusively through the mobile app. Google plans to add more features during the preview, but for now, it primarily functions as a conversational AI guide, helping with fitness and sleep while identifying trends in your health data. The app features four main tabs: Today, highlighting timely insights; Fitness, for creating a weekly workout plan; Sleep, offering a detailed look at your nightly patterns; and Health, which tracks your overall metrics.
I’ve thoroughly tested the personal health coach, and after five weeks, I’m ready to call it the most effective automated health coach I’ve used—though it’s clear there’s room for improvement. Here’s a detailed look at my experience.
Getting Started: Onboarding That Feels Like a Real Trainer
To get started with the Google personal health coach, you must be logged in to the Fitbit app with your Google account. During the preview, the coach is only available in the US in English for users aged 18 years or older. Note that not all Fitbit features have rolled forward to the new personal health coach interface yet, but you can always revert to the old app if you don’t like the new experience.
(Credit: Google/PCMag)
As mentioned, the experience requires a Fitbit Premium membership, and you’ll also need a Fitbit device that supports the brand’s Cardio Load metric. Your fitness tracker options include the following Fitbit wearables: the Inspire 2 or 3, the Luxe, the Sense or Sense 2, the Versa 2 through 4, or the Charge 5 or 6. For smartwatches, any of Google’s Pixel Watch models will work. I used the Pixel Watch 4.
If you have all the prerequisites, a card will appear in the app offering the option to update. I clicked the card, agreed to give the AI access to my data, and granted Google permission to use my data for research (the first permission is mandatory to use the coach, the second is optional).
Next, the AI coach interviews you to understand your goals and history and to set up an appropriate health plan; Google estimates that this step takes around 10 minutes in total. You can skip the interview and come back to it later, but I wanted to dive right in. During this interview, I conversed with the AI via a text interface in the app. The conversation reminded me of the initial chats I had with my real-life personal trainer, Jorge (an actual human, not AI), whom I see twice a week.
Snippets from my onboarding conversation (Credit: Google/PCMag)
The coach asked me about my main goal, what I thought was working, and my biggest challenge in pursuit of that objective. It covers the basics of using the program and highlights the Ask Coach button, which is available in all app tabs. The initial conversation lasted approximately five minutes.
Creating the fitness plan took a few minutes longer, but I had some specific requests. At the outset, you can set different fitness goals from the ones you created for your overall health if you’d like. I’d been wearing the Pixel Watch 4 for a couple of weeks before onboarding, so it had some workout and sleep data to use for contextual information. It used that information to highlight the patterns it had observed, including my past activities, which it used as a basis for exercise recommendations, and my sporadic workout habits, which it took into account when setting goals.
It then asked what equipment I have access to, while pointing out that I likely have a heart rate monitor (which seems self-explanatory, given that a fitness tracker is required), as well as gym equipment for my past high-intensity interval training (HIIT) sessions. I confirmed its assumptions, then got specific regarding my own needs.
I specified that I only have access to gym equipment during my two personal training sessions a week. I told the AI that I don’t need help with those workouts, but could use some motivation to exercise more throughout the week.
The coach factored in that information and recommended a plan with four workouts a week, including my two personal training sessions. It also switched from recommending a bodyweight strength workout to a bodyweight cardio workout, as I focus on strength with my trainer. Additionally, it recommended a run for the other session, given that I’ve logged runs in the past.
Once adjusted, the coach asked for 10 minutes to process all the information, but my plan was finished and ready for my perusal within roughly 90 seconds. Everything matched what I specified in the chat, and I was excited to get started.
While the setup took some time, given that the whole thing takes place in a text chat, it was easy to do and highly responsive to all my personal requests and needs, while also taking my workout experience into account.
Personalized Workouts: A Plan That Adapts to Your Life
Once up and running with Google’s personal health coach, it would display its proposed plan for the following week each Saturday evening. However, I could use a button to make changes at any time via a chat window similar to the onboarding conversation.
I started with the AI coach mid-week, and since it hadn’t fully figured out my schedule with my personal trainer, Jorge, yet, I had to rearrange everything in the second week, which was easy enough. I informed the AI about the scheduled training sessions and my availability, and it made the necessary adjustments. On subsequent weeks, it learned my schedule and planned accordingly.
Whether planning a week of workouts or making tweaks on the fly, the coach always proved reliably fluid. It once prescribed a run on a day when it was too cold, so I asked for a substitute, and it produced a circuit of bodyweight cardio exercises. Google notes it can pull from a library of over 700 different movements depending on the equipment you have on hand and your stated goals.
As I used the app, it continued to refine its recommendations. It gradually increased the intensity of its recommended workouts and offered compliments and words of encouragement when I stayed on schedule.
(Credit: Google/PCMag)
The coach even proactively suggested changes when it noticed my unusual Thanksgiving schedule. My family had to stay with me due to winter storms in Chicago, and the AI detected an elevated heart rate and lower quality of sleep. It suggested I alter my planned run from a long, intense session to a short, recovery-focused walk. I gladly accepted. With other guides, I’d have just skipped the workout altogether. The coach is smart enough to recognize that something is better than nothing and kept me moving with a suitable compromise.
In the Fitness tab, it shows your remaining weekly plan, and tapping each workout pulls up details and advice. Due to its fluidity, I was continually motivated to follow the plan as closely as possible and keep the AI updated when adjustments were needed.
Get Our Best Stories!
Your Daily Dose of Our Top Tech News
Thanks for signing up!
Your subscription has been confirmed. Keep an eye on your inbox!
The coach takes into account your Daily Readiness Score (a metric that indicates how hard to push that day, based on your recent workouts and recovery status) and your Cardio Load value (a measure of total strain on your body from activity and exercise). The coach uses Cardio Load as a weekly goal, whereas the Fitbit app introduced it as a daily metric.
During my testing, I consistently met this Cardio Load target and observed a steady increase in my Daily Readiness Score as a result. I also enjoyed seeing my Cardio Load progress leap forward on the wheel at the top of the Today tab after a tough workout.
Small Glitches, Big Potential: Usability Hiccups to Watch
With some exercises, like running, the app offers a button to send the workout to my watch for live tracking. With others, I have to manually start a workout on my watch, and then link the tracked exercise to the app’s recommendation afterwards. While not a huge issue, I would prefer to be able to send all types of workouts to my watch via the app.
The coach has a few other usability hiccups. Notifications regularly pop up with workout summaries, recaps of daily activity, sleep data, and more. They aren’t so frequent as to be intrusive, and I enjoy viewing the information within the notification. That said, sometimes the entire text doesn’t fit. Tapping the notification pulls up the app, but it doesn’t always bring you to the same information to finish reading it.
After a workout, for instance, the coach will send you a summary filled with compliments about why it was a smart choice. Tapping this notification takes you directly to the actual stat breakdown of the workout, rather than the remainder of the summary. To find the summary, you need to close the workout stats and return to the main Today tab, then scroll down.
When it suggested a reduced workout after Thanksgiving, I accepted the recommendation via a pop-up card in the Today tab, but then the text interface wasn’t sure what I was talking about, and I had to clarify. It only took a moment, but it was a strange disconnect.
I’d also appreciate a deeper integration between Fitbit’s workout library and the coach. For one workout, it recommended a 25-minute yoga flow session. It wanted to accommodate my stated goal of trying yoga and increasing flexibility. The recommendation made sense, but when I went to do the workout one Saturday, it offered no guidance or instruction. When I chatted with the coach about offering some guidance, it said it had none to give and switched my workout to a bodyweight cardio session so it could offer activities from its list, like burpees and mountain climbers.
Recommended by Our Editors
(Credit: Google/PCMag)
Given that Google has access to Fitbit’s library of workout videos, this problem feels easily solvable. Even beyond that, Google owns YouTube, which is loaded with free guided exercise videos.
Even when the coach has suggestions, they don’t always match the overall recommendation. For one session, it recommended a 30-minute workout, but only provided enough exercises to fill 10 minutes. The suggested workout already involved several sets of the same exercises. Perhaps it is doing its best, given my lack of equipment at home, but it’s odd that the timing and intensity of its suggested activities don’t always align with the overall goal of the workout.
Using it for a running workout was more seamless, but slightly annoying. It set different goals for my weekly runs, with some focused on pace, others on heart rate, and others on distance. During the heart rate run, the coach caused my watch to chime too frequently with information indicating that I was slightly above or slightly below the target. The problem is unintuitively alleviated by being way off the mark, as it only speaks up when you are just entering or exiting range. Thus, on runs when I tried to follow its recommendations closely, it annoyed me the most.
Finally, this might be strictly a personal problem, but I sometimes found the coach to be a little too complimentary. It only had nice things to say about any of my workouts, including those in which I was distinctly outside the goal range. My personal trainer, Jorge, knows that I respond well to a stern push on occasion. Google’s AI coach operates from the understandable and welcome assumption that something is better than nothing when it comes to working out, but small critiques here and there, or even just advice on form when my device can track such qualities, would be a welcome addition.
Sleep and Health Tracking: Insightful, But Limited
Despite the room for improvement, I’m impressed with the personal health coach from a fitness standpoint, but I haven’t noticed that much difference between it and the basic Fitbit app on the other two tabs. Google has acknowledged that sleep, in particular, is a work in progress. Promised features include recognition of naps and restful periods throughout the day, as well as more detailed recommendations for improving sleep quality.
In testing, the coach assessed my sleep and noted if I got too little or deviated from my usual schedule. It scored my shut-eye, and the Today tab showed summaries discussing the quality and duration of my deep sleep. The coach was accurate here, but mostly just offered text descriptions of information I could already see from the graphs in the old Fitbit app.
The Health tab deliberately does not give medical recommendations, so the coach is just there to spot trends, make observations, and answer questions. On that front, it paid off with my Thanksgiving workout adjustment when it noticed a spike in my resting heart rate.
Otherwise, the Ask Coach button is present in these tabs as well, and you can use it to chat about any tracked metric or even about how you slept and why you slept that way.
Verdict: It’s the Closest Thing to an Actual Coach I’ve Tried
Google wearables aren’t your only option for this type of AI coaching. Samsung’s latest Galaxy Watches launched with both a Running Coach and a Sleep Coach powered by Galaxy AI. Samsung’s Sleep Coach offers a bit more advice than the Google personal health coach about steps to take before bed to get a better night’s sleep. The Running Coach asks you to take an assessment test and then creates a program tailored to train for an event, such as a 5K, using your initial results as a baseline. In practice, it offers similar goal guidance to Google’s running workouts and is similarly insightful, although it can also be a bit overwhelming when approaching a goal threshold.
Apple’s Workout Buddy can be overbearing as well when you activate goal metrics. While not designed to be prescriptive on its own, Workout Buddy takes into account your established pace or heart rate goals, and will chime in a little too frequently if you’re bouncing above and below a target threshold. Otherwise, it’s the simplest of the three, as it’s mostly meant to provide encouraging words and stat updates as you move. With eight supported activities, Apple’s AI works for a wider range of exercises than Samsung’s Running Buddy, which is limited to one.
Samsung currently has the edge at offering sleep advice; neither Samsung nor Apple’s workout guidance comes close to the adaptable and generative nature of Google’s personal health coach. Interacting with the coach is as easy as sending a text. It allows you to make adjustments, ask questions, and get advice, all from this same interface.
Polar and Garmin offer long-term workout plans in their respective apps, but these apps lack the interactivity of Google’s personal health coach. Aside from popular wearable apps, you can generate workout advice through a general AI like ChatGPT or Google’s own Gemini, but the health coach uses the same engine as the latter and takes into account the data from wearables in a way those apps can’t.
While not perfect, the Google personal health coach is ahead of the pack. Some of the necessary refinements may emerge as the software approaches its final launch. Even as it stands, it’s interactive enough that I felt some accountability to stick to its schedule, knowing that it would accommodate me when conflicts did arise. It worked well with my existing workouts and customized a plan based on my needs, and I’m now more fit than I was five weeks ago when I started testing it.
About Our Expert
Andrew Gebhart
Senior Writer, Smart Home and Wearables
Experience
I’m PCMag’s senior writer covering smart home and wearable devices. I’ve been reporting on tech professionally for nearly a decade and have been obsessing about it for much longer than that. Prior to joining PCMag, I made educational videos for an electronics store called Abt Electronics in Illinois, and before that, I spent eight years covering the smart home market for CNET.
I foster many flavors of nerdom in my personal life. I’m an avid board gamer and video gamer. I love fantasy football, which I view as a combination of role-playing games and sports. Plus, I can talk to you about craft beer for hours and am on a personal quest to have a flight of beer at each microbrewery in my home city of Chicago.
Areas of Expertise
Latest By Andrew Gebhart