Google’s AI-powered Health Coach in the Fitbit Google Health app was tested publicly with opt-in users, and one early adopter says it delivered too many notifications, missed context during travel, and offered repetitive advice instead of genuinely useful insi
Google has teased its new AI-powered Health Coach in the Fitbit Google Health app for months. It also began testing the feature publicly with users who opted into the experiment — and for one reviewer, the rollout couldn’t have come at a more inconvenient time.
A few weeks ago, just before a week-long trip, a notification appeared asking whether they wanted to try the new health coach. They agreed immediately and, over the first week, moved through Romania while watching what the coach said about their sleep and activity.
At first, it felt almost harmless — like listening to someone who genuinely cares. But the pattern was relentless. When they woke up, the coach appeared to comment on their sleep. When they walked from a hotel to downtown Bucharest, the app sent another message. Another walk after lunch triggered yet another. Then, on another trip back to the hotel after grabbing a coffee, the coach chimed in again.
The more they moved — especially because travel meant frequent stops, bursts of walking, and less predictable routines — the more notifications arrived. What started as amusing quickly turned into noise.
They said the coach’s updates were arriving even after short strolls that didn’t “move the needle” for their health. In their view, the coach should understand that not every small movement during an active day deserves interruption. They argue that if someone has been sedentary most of the day and then takes a quick 15-minute walk. that’s one thing. If someone is already moving throughout the day. they say only something more substantial — longer than 30 minutes — should reliably trigger a notification.
It wasn’t just the frequency. They also criticized the quality of the “insight” being delivered. Messages seemed to boil down to whether they walked at a brisk pace or an easy pace. and whether their heart rate rose “too much” or “not too much.” But their experience was complicated: they were strolling. or they were going uphill while dealing with asthma. which they say affects what they can realistically expect from heart-rate changes.
By notifying them about every little activity, the Health Coach started to condition them to ignore its feedback. Too much noise. they wrote. meant less signal — and by the time the trip ended. they said they were almost trained not to look at the coach at all because it talked constantly without saying much that felt truly meaningful.
Their frustration sharpened further when they described how the coach failed to adapt after they responded with context. They said they tried to reply to the AI and explain what travel and real-life limitations were doing to their routine. But they felt the coach “didn’t adapt all that much” to those answers. After a couple of walks. hikes. or nights of bad sleep. it reportedly returned to the same nagging suggestions about doing things right — the coach’s version of right.
They described the advice as repetitive and increasingly irritating: turn off screens and rest earlier tonight, you’ve done more than enough today, dim the lights and give your brain a rest, and it’s time to call it a day.
They also acknowledged that not every bad week comes from poor choices. They said they had “terrible nights of sleep” over the last month or so alongside “a lot of active days. ” and that Google’s Health Coach didn’t appreciate that combination. Sometimes. they argued. people don’t get to choose their conditions — travel. or sleep disruption tied to anxiety issues — and an app coach should respond with more restraint.
In their view, a truly intelligent coach would know when to shut up. They said the Health Coach interpreted every small thing it detected and treated every detected behavior as an opportunity to comment. They even compared it to a mom who wants the best for them, but doesn’t know when to stop.
Their complaint comes down to control and timing: Google should let users tailor how much feedback they get. and possibly adjust it over time — more feedback during weeks they’re receptive. less during weeks they want to be left alone. They also framed the expectation differently than an in-person coach. If someone paid for a human coach, they said constant nagging might make sense. But on a phone. they argue. it should follow the user’s cues and tone itself down when it’s clearly not being heeded.
They ended with a simple hope: for the coach to adapt — both in how often it interrupts and in how much it repeats — so the guidance feels more contextual, and the important signals don’t disappear inside a flood of notifications.
Fitbit Google Health AI Health Coach smartwatch health health app sleep insights user feedback notifications