Wearables as Coaches: How AI is Moving Sport Tech From Recording to Recommending
- Shane Riddle
- 4 hours ago
- 14 min read
For most of the last decade, a fitness wearable answered one question: “What did I do?” Step counts, calories, heart rate during that Tuesday run. Fine, but not coaching. The question has shifted to “What should I do next?” and that shift is why the American College of Sports Medicine named wearable technology the number one global fitness trend for 2026, based on responses from around 2,000 fitness and exercise professionals. This post breaks down what's actually available now, what's genuinely in development, and what still lives on a product roadmap, with no hype attached — a wrist-level companion to our broader look at AI in sport.
Key Takeaways
ACSM's top trend for 2026: Wearable technology topped the ACSM's annual global fitness trends survey for 2026, with nearly half of U.S. adults now owning a tracker or smartwatch.
The shift is from recording to recommending: Devices like WHOOP, Garmin, and Oura now use HRV, sleep quality, and recovery data to suggest what workout you should do — or skip — today. That's a different job than logging a step count.
WHOOP Coach is the clearest example of AI coaching that's available right now: Powered by OpenAI's GPT-4 and trained on your personal health data, it offers genuinely individualised conversational coaching. You can ask it why your recovery was low and get an answer tied to your specific metrics.
The science behind HRV-guided training is solid: Multiple peer-reviewed meta-analyses support HRV-guided training as a way to improve VO2 max and reduce overtraining risk compared to fixed training schedules.
Know what's here vs. what's coming: Adaptive AI that auto-adjusts your plan in the background without you asking is still in development. Non-invasive blood glucose monitoring in a mainstream smartwatch doesn't exist yet. Both are worth watching.
Standards bodies are catching up, slowly: IEEE and ACSM are working to define what "accurate" and "good" looks like for wearable health data. Consumer devices aren't required to meet a single unified standard yet, so independent validation studies matter more than manufacturer claims.
Table of Contents
Wearables Quick Facts
Wearable technology was named the #1 fitness trend for 2026 by the ACSM, based on a survey of approximately 2,000 fitness professionals. Source: ACSM
Nearly half of U.S. adults now own a fitness tracker or smartwatch. Source: ACSM 2026 Trends
The global wearable technology market sits at roughly $90-100 billion in 2025, projected to exceed $229 billion by 2033 at a CAGR of around 12%. Source: Grand View Research
Sources: ACSM, Grand View Research
The Shift From Recording to Recommending
Ten years ago, a good fitness tracker told you how many steps you took. Five years ago it told you how well you slept. Right now, the better ones are starting to tell you what to do about both.
There's a real change as AI now takes hold across these devices. The sensors are getting better, but more importantly the algorithms changed. Early wearables worked with population averages, your heart rate during sleep, compared against millions of other users. Useful for spotting oddities, but not coaching, because it doesn't know you, specifically, over time.
What the current generation of AI coaching tools are doing is accumulate enough data about your own patterns, your resting heart rate baseline, your Heart Rate Variability (HRV), your sleep debt, your recent training load to give advice calibrated to you rather than to a statistical average. There is still a gap with this tech, a coach who's worked with you for six months knows things a wearable can't measure for example your stress at work, your motivation on a cold morning, your tendency to overcook intervals. But this gap is narrowing as AI continues to develop in this area.
What's Actually Available Today
WHOOP Coach
WHOOP launched WHOOP Coach, powered by OpenAI's GPT-4, to all members. You can ask it questions about your own data like, "Why was my recovery low this week?" "Should I push hard today or hold back?" It gives answers based on your actual metrics, not generic advice.
WHOOP 5.0 also introduced a medical-grade tier (WHOOP MG) available on a yearly subscription, which includes on-demand ECG and blood pressure insights.
Garmin Daily Suggested Workouts
Garmin's Daily Suggested Workouts are already on most modern Garmin devices. The algorithm pulls your training load from the past four weeks, your current recovery time, and your VO2 max estimate to recommend what type of session to do today and at what intensity.
Garmin Coach goes a step further: pick a goal race distance, choose a coaching profile (Greg McMillan, Amy Parkerson-Mitchell, or Jeff Galloway), and get a training plan that adjusts week to week based on how you're actually running.
Oura Advisor
Oura Advisor is an LLM-powered assistant built into the Oura Ring app. In April 2026, Oura released a proprietary AI model for women's health within Advisor, built on medical literature and reviewed by board-certified clinicians, integrating temperature sensing with cycle prediction and fertility window estimation.
The Readiness Score itself, a 0-100 figure calculated from sleep quality, HRV, resting heart rate, and activity balance tells you each morning whether you're primed for effort or better served by recovery. Oura's guidance is that scores above 85 suggest you're ready to push, while scores below 70 are a signal to back off.
Apple Watch Training Load
watchOS 11, available from September 2024, introduced Training Load to Apple Watch, comparing your workout effort over the past seven days against your 28-day average. A year later, watchOS 26 added Workout Buddy, what Apple calls a “first-of-its-kind workout experience with Apple Intelligence” which draws on your workout history and live metrics to deliver personalised spoken motivation during sessions. It runs on Apple Watch Series 6 and later, SE 2 and later, and all Ultra models, and needs an Apple Intelligence-enabled iPhone nearby plus Bluetooth headphones.
One caveat worth knowing, Apple Watch measures HRV at spot intervals during sleep rather than continuously, which limits its overnight recovery precision compared to WHOOP and Oura.
How the Leading Devices Compare
WHOOP 5.0 / MG — Screenless strap with 14+ days of battery. WHOOP Coach offers conversational, data-grounded AI coaching. Continuous overnight HRV. Best for recovery-focused athletes comfortable with a yearly subscription that has the hardware included.
Oura Ring Gen 4 — A ring with roughly 8 days of battery. Oura Advisor provides LLM-based guidance, now including a dedicated women's health model. Continuous overnight HRV with the strongest independent validation of the devices here. Best for sleep-first tracking and discreet wear. There is an upfront cost for the ring and then an ongoing monthly membership.
Garmin Fenix 8 — Multisport smartwatch with up to 29 days of battery. Daily Suggested Workouts and Garmin Coach adapt training plans week to week. Overnight readings feed a rolling HRV Status each morning. Best for GPS endurance athletes and has an upfront cost for hardware and no no subscription.
Apple Watch Series 10 / Ultra 2 — Smartwatch with 18 hours (Series 10) to 60 hours (Ultra 2, low power) of battery. Training Load from watchOS 11, and Workout Buddy from watchOS 26 on supported models. HRV is spot-checked during sleep rather than continuous. Best for general smartwatch users in the Apple ecosystem and requires the purchase of the smartwatch with no additional subscription.
How much should you trust these numbers? Please do your own research on these as models and features continue to change. As a guide our research found an independent 2025 validation study published in Physiological Reports which compared five devices against an ECG reference across 536 nights of real-world sleep, and found the Oura Ring Gen 3 and Gen 4 showed the highest agreement for both overnight HRV and resting heart rate, with WHOOP 4.0 showing acceptable agreement and the Garmin Fenix 6 and Polar trailing. One caveat of the study, it tested the previous WHOOP generation and an older Garmin watch, so read it as a guide to each platform's sensing approach rather than a verdict on the current models.
Can a Wearable Replace a Human Coach?
It's the question the whole premise begs, so let's unpack it properly. It depends on which part of coaching you mean, because the two are good at almost entirely different things.
Cost: The subscriptions in this post run from nothing at all (Garmin) and at time of writing, $359 a year (WHOOP's top tier). A human coach charges per session, and even one session a week will pass the annual cost of any device here within a few months.
Availability: The wearable is on duty every night and every session, including the 2am ones. A human works in scheduled hours, and the best ones have waiting lists.
What it can see: The wearable measures things no human can observe, your heart rhythm through the night, every night, for years. The human sees things no wrist sensor measures, your technique, your movement quality, and the look on your face two reps before you'd admit anything is wrong.
Context: You can tell WHOOP Coach you're under stress at work, and it will factor that in, if you remember to say so. A good human coach notices without being told.
Accountability: A notification is easy to swipe away. A person expecting you at six in the morning is not.
The conclusion I come to is that they aren't competing for the same job. The wearable is the most complete physiological record you will ever own, the human is the better judge of what to do with it. If you can only have one and you're chasing real improvement, take the human. But the strongest setup isn't either–or, it's a coach who can see your recovery data and a device that fills in the nights and weeks between conversations.
The Risks: What an AI Coach Won't Catch
None of this is a reason to bin your strap or monitor, but we need to understand any potential risks with this tech. Hunter Bennett, an exercise scientist at Adelaide University, examined the safety question in a July 2026 piece for The Conversation, and his concerns map neatly onto wearable coaching. Here's what the research and common sense says can go wrong.
Nobody screened you first: An accredited exercise professional checks for injuries and health conditions before writing you a program as a basic safety step. Bennett points out it isn't clear that AI tools can properly account for injuries or medical conditions, which means their recommendations may simply not be safe for your current health. If you have a chronic condition or a history of injury, that gap matters more than any feature on the spec sheet.
It can green-light a session your body isn't ready for: Injury risk climbs when training volume or intensity jumps too quickly, and Bennett notes AI-generated guidance may not factor that in. A wearable adds a specific version of this problem, a high recovery score reads as permission, but the algorithm can't know your calf has been grumbling for a fortnight. Recovered physiology is not the same thing as a body that's ready.
It has never watched you move: The riskiest place for confident advice is a new movement performed badly, and no wearable sees your technique. Bennett's practical suggestion for anyone new to the gym is worth repeating, book a couple of sessions with a human first to learn the movements, then let the AI handle the programming. This one resonates as it's a practice we actively promote at the recreation centres YMCA Victoria manages, where I work, start with a qualified instructor on the gym floor, then bring the technology in.
It only knows what you tell it: The research Bennett reviews found AI program quality depends heavily on the detail you provide and knowing what detail matters is itself expertise most beginners don't have. WHOOP Coach will factor in your work stress if you mention it. If you don't, it's coaching a calmer person than the one wearing the strap.
The head-to-head evidence still favours humans: The comparison studies Bennett cites point the same way, a twelve-week trial found bigger strength and muscle gains under a personal trainer than under ChatGPT's guidance, and shorter studies found modest but consistent advantages for human-written programs. The differences aren't dramatic but the direction is.
A daily score can become a daily judgement: This one is ours rather than Bennett's: some people find a readiness number motivating, and some find it quietly corrosive, checking it before deciding how to feel about the day. If that's you, it isn't a discipline problem to fix. It's a signal the tool costs you more than it gives.
The sensible reading of all this is AI coaching is well suited to the broad middle, healthy adults doing ordinary training who want better-informed decisions. At the edges, Bennett's conclusion is the right one. If you're injured, managing a chronic condition, brand new to training, or chasing genuinely high performance, the safest and most effective option is still a qualified human.
What's Still in the Pipeline
WHOOP's Adaptive Coaching — in development: WHOOP has published Adaptive Coaching on its product roadmap, a background system using real-time learning and predictive AI models to adjust guidance for training, recovery and lifestyle habits without requiring input from you. As of mid-2026 it remains a roadmap feature rather than something general members can use.
Continuous blood glucose monitoring in mainstream wearables — in development:
Several companies are working on non-invasive or minimally invasive continuous glucose monitoring for sport wearables. Abbott's Lingo device is available as a consumer product in some markets, but it's a separate adhesive patch worn on the arm, not integrated into a smartwatch. As of mid-2026, no current Apple Watch model includes non-invasive glucose sensing. For endurance athletes, knowing blood glucose in real time during training would change fuelling strategy in ways that matter.
Fully on-device AI with local learning — research phase: Current wearable AI coaching depends heavily on the cloud. Researchers are exploring models small enough to run meaningful personalisation on the device itself, improving privacy and reducing latency. Practical implementation in a consumer wearable is probably several product cycles away.
Standards Bodies and Governance
ACSM: The American College of Sports Medicine is US-based with global influence with exercise-science standards, the annual fitness trends survey, and guidance for clinicians on interpreting wearable data.
IEEE: The IEEE publishes international technical standards for wearable consumer devices, covering hardware and measurement.
Consumer Technology Association: The CTA maintains voluntary US accuracy standards for step, heart rate and sleep tracking.
The practical upshot for anyone buying a wearable today is the "accuracy" claims on a product should be treated skeptically until you find an independent validation study. A useful independent resource is the blog DC Rainmaker, which publishes detailed accuracy comparisons for most major devices.
Do Wearables Coach Every Body?
Mainstream fitness algorithms are overwhelmingly built and validated on able-bodied users. A step count means nothing to a wheelchair user, and recovery models tuned to running and lifting loads don't automatically translate to how a para athlete actually trains.
Some of it is being addressed, and it's available today. Apple Watch includes a dedicated wheelchair setting that tracks pushes rather than steps, registers different push types, speeds and terrains, swaps the Stand goal for a Roll goal, and adds two dedicated outdoor push workouts. It's a genuine example of a mainstream wearable designing for a different body rather than retrofitting one.
The AI coaching layer hasn't caught up yet. None of the conversational coaches covered here advertises training guidance designed for para athletes, so for athletes in sports like wheelchair basketball, the raw signals, sleep, heart rate, HRV remain useful, but the coaching advice is calibrated for someone else. That's worth knowing before you lean on the recommendations, and it's an area where the industry has real work to do.
Wearables for Children and Teens
The recovery-focused platforms in this post are adult products. Oura states plainly that its ring is not intended for anyone under 18. WHOOP's terms of use set minimum ages for holding an account up to 18 in some jurisdictions and several features are age-restricted regardless. Healthspan isn't available under 18, and the ECG feature isn't intended for anyone under 22.
This isn't gatekeeping for its own sake. These platforms are designed and validated for adults, and a teenager's sleep architecture and heart-rate patterns are still developing. Therefore a readiness score tuned to adults isn't reading a young body accurately. There's also a fair question about whether a daily number judging a child's “recovery” is a healthy thing to hand them at all.
What does exist for children is deliberately different. Garmin's Bounce is a kids' smartwatch built around activity, messaging and location rather than recovery metrics. Apple Watch For Your Kids (formerly Family Setup) gives children activity rings with move minutes in place of calorie targets, and deliberately leaves out features like sleep staging and blood oxygen. The design philosophy encourages movement and skips the physiological scoring which is the right one for young athletes.
One more thing before you hand any tracker to a child. My working life sits between technology and community sport, and this is the part I care most about, a child's biometric data deserves more protection than an adult's, not less. Who owns and controls athlete body data and what that means in junior sport is exactly what our guide to athlete biometric data rights covers.
Final Thoughts
The wearables conversation used to be simple, your device recorded what happened, you looked at it, you maybe changed something. That loop is getting a lot more active. WHOOP Coach, Oura Advisor, and Garmin's Daily Suggested Workouts represent real functionality, available today, that tries to close the gap between data and decision.
It's not magic. It doesn't replace human judgment or a good coach who knows you well. But for athletes who train without a dedicated coaching relationship which is most people, a system that reads your overnight HRV, cross-references your recent training load, and says "this is a recovery day" is a step change from a step count.
The things worth watching over the next two or three years are adaptive AI that updates your training without prompting, better integration between sleep staging and workout planning, and the eventual arrival of glucose monitoring in a mainstream form factor. For now, the honest framing is this: understand what your device does well and where it's still guessing. Whether it translates into better training is still, ultimately, up to you.
Frequently Asked Questions
Q: Is HRV-guided training actually better than following a fixed training plan?
A: The research says yes, at least for improving VO2 max and aerobic fitness. Multiple peer-reviewed studies, including a meta-analysis in PMC, found HRV-guided training produced better outcomes than pre-set schedules. The caveat is that the quality of the HRV data matters, and so does your understanding of what's driving the numbers on any given day.
Q: What's the difference between WHOOP Coach and what Garmin or Apple does?
A: WHOOP Coach is a conversational AI, powered by GPT-4, that responds to your personal data and lets you ask specific questions. Garmin's Daily Suggested Workouts is algorithmic, it outputs a recommendation based on your training load and recovery, but there's no conversation. Apple Watch Training Load is more of a status indicator than a coaching tool. They're doing related but different things.
Q: Can I use Oura Ring as my main training tracker?
A: It depends on what you track. Oura Ring Gen 4 is one of the most accurate consumer devices for overnight HRV and resting heart rate, and its Readiness Score is a solid input for training decisions. It doesn't have GPS and doesn't do sport-specific tracking the way Garmin does, so many athletes pair it with a GPS watch rather than using it alone.
Q: What's WHOOP's Adaptive Coaching feature and when is it coming?
A: Adaptive Coaching would automatically adjust your training guidance in real time based on your recovery, sleep, and lifestyle signals, without requiring you to ask. As of mid-2026, WHOOP has confirmed it's on the roadmap but it's not yet available to general members.
Q: How do I know if my wearable's data is accurate enough to trust for training decisions?
A: Look for independent validation studies, not manufacturer claims. For overnight HRV and resting heart rate, a 2025 study in Physiological Reports found the Oura Ring Gen 3 and Gen 4 showed the strongest agreement with an ECG reference among the consumer devices tested. DC Rainmaker publishes detailed, independent accuracy comparisons for most major devices and is a reliable resource.
Q: Can my child or teenager use an AI coaching wearable?
A: The recovery platforms are adult products, Oura's ring is not intended for under-18s, and WHOOP age-restricts accounts and features such as Healthspan (18+) and ECG (22+). For children, purpose-built options like Garmin Bounce or Apple Watch For Your Kids focus on movement and safety rather than recovery scores, which suits developing bodies far better.
Related Posts:
