A fitness app gives you a plan, a place to log every set and a reminder to train. An AI chatbot writes a draft plan from whatever you type into it. An online coach writes the plan around your week, reads your numbers every week and changes the plan with you. The first two are good at the plan. The third is about what happens after the plan, on the weeks it stops working.
I'm an online coach, so I have a stake in this. I'll try to be fair, including about when an app or a chatbot is all you need, and about how little the research can settle.
Side by side
This compares three kinds of help, not particular products. Apps and AI tools differ a lot, so treat those two columns as what to check. The coach column is how I work.
| Fitness app | AI chatbot | Online coach (how I work) | |
|---|---|---|---|
| What you get | A plan, a place to log your sessions, reminders and streaks. | A written plan, drafted from your prompt. | A plan built around your goal, schedule and equipment, in the Trainerize app, and a person who checks in. |
| Who writes the plan | Software, from your answers. | Software, from what you type. | I do. |
| Who changes it | The app's rules, from what you log. | You, by asking again. | I do, with you, on the weekly call. |
| Who reads your numbers | The software. | Only what you paste into the chat. | Me: two weigh-ins a week, progress photos every two weeks and your logged workouts. |
| Technique | Varies by app. | Varies by tool. | Taught as you go; send a video of a lift if you want notes back. Not live. |
| Cost in time | Your sessions, plus logging them. | Your sessions, plus writing a detailed prompt and checking the answer. | Your sessions, plus two weigh-ins a week, photos every two weeks and a weekly call. |
| Cost in money | Varies by app. Check what a free version leaves out before you pay. | Varies by tool and plan. Check the maker's own pricing page. | I don't publish a price list. It depends on the support you need, and we talk about it first, in the DMs. |
| Suits you if | You know what to do and mostly need a place to log it and a nudge. | You want a starting draft and can judge whether a plan makes sense. | You keep starting over, your week keeps changing, or what's missing is someone reading your numbers. |
| Falls short when | Your week changes and nobody changes the plan with you. | The details you left out matter, or something needs a professional. | You want someone beside you every session, or you already train consistently on your own. |
What a fitness app gives you, and where it stops
An app on its own can give you a plan, a place to log every set and a reminder to train. That's not nothing. A 2021 review pooled 28 randomized trials of apps and activity trackers in 7,454 adults aged 18 to 65 without chronic illness, and found a small-to-moderate increase in physical activity, about 1,850 steps a day. Apps that included text messages and personalization did better.1
For lifting, the evidence is thinner. A 2025 review of 32 randomized trials looked at strength training prescribed through an app or a website. Against no program or usual care, it found a small improvement in strength and related measures: a moderate effect for the lower body and no significant effect for the upper body. Most of the trials were in people with a health condition, and most used bodyweight exercises.2
Where it stops: an app reacts to what you log. It doesn't know why you stopped: that you're travelling next week, or that work blew up. And a little human contact bolted onto an app isn't automatically the fix. In a 2020 trial, 357 US veterans (average age about 40) used an activity app for a year; half also got weekly goals, three automated motivational messages a week and up to three coaching phone calls. Activity went down in both groups, with no difference between them at six or twelve months, and many people had stopped recording their activity by the end, so the result is uncertain.3 Three calls in a year is not weekly coaching, so this trial can't say whether regular contact would help. It does say that a little contact bolted onto an app isn't automatically the fix.
What an AI chatbot gives you, and where it stops
A chatbot will write you a training plan from a paragraph. Researchers have started grading those plans, and the pattern so far: a usable draft, not a finished program.
In a 2024 study, researchers asked ChatGPT (two versions) to design 12-week resistance training programs for hypothetical 20-year-olds with intermediate experience, then compared the programs with established guidelines. It produced structured plans in three four-week phases, and the programs were similar for men and women. The authors concluded the programs likely needed further fine-tuning, and that the tool may serve for writing a draft program but likely needs human expertise to make the program as effective as it can be.4
The prompt matters. In a 2025 study, 12 coaching experts rated eight muscle-building plans written by two AI chatbots. Giving the same prompt again produced plans of similar quality (27 of 28 items showed no difference), and prompts with more detail produced plans that the experts rated higher.5 So a vague prompt tends to get a lower-rated plan.
A third study, also from 2024, asked a chatbot for a 30-day program for each of five hypothetical people with different health conditions and goals, from a single first exchange, much like a typical user would. Experts found it could write general, safety-conscious programs, but the programs lacked precision for each person's condition and goals, often putting excessive safety ahead of training effectiveness. The authors said these tools can be a supplement, but aren't yet recommended as a substitute for personalized, progressive prescriptions from health and fitness professionals.6
One tool maker says something similar. Google's terms of service, which apply to its AI services too, make no promise that what its services produce is accurate, and say not to rely on them for medical, legal, financial or other professional advice.7
Where it stops: in all three studies, the plans were judged on paper, by the researchers or by expert panels. Nobody trained on them, so none of it tells you what results a chatbot's plan gets. And a chatbot only knows what you tell it. It doesn't see the week you actually had unless you type it in.
What an online coach gives you, and where it stops
With me, your workouts live in the Trainerize app. You weigh in twice a week and send progress photos every two weeks. We talk on a weekly call scheduled around your time zone, and you can message me with questions between calls. On the call we look at the week you actually had and change the plan for the one that's coming.
So the plan is the smaller part. An app has a plan too; mine even lives in one. What a coach adds is the reading: someone who looks at your numbers every week and asks what happened. There's a reason that might matter. A meta-analysis of 138 randomized studies (19,951 people) found that getting people to monitor their progress toward a goal helped them reach it, and the effect was larger when progress was reported or made public, and when it was physically recorded.8 Those studies aren't about coaching. A weekly check-in is a version of both.
One trial put the three formats closest to this question side by side. In a 2025 randomized trial, 112 trained adults started the same 10-week, three-days-a-week program with a coach beside them, guided by an app (the authors call it online coaching), or from a PDF. Of those who started, 88% of the group with a coach beside them, 81% of the app group and 52% of the PDF group completed at least 26 of the 30 sessions. Every group that finished got stronger, and the coached group gained more on the squat than the other two. The authors concluded that training with a coach beside you gave the best results overall, and that the app version shows promise for keeping people training when full supervision isn't possible.9 There was no AI group, and no group had a weekly call with a remote coach, so it can't rank coaching like mine against an app.
Where it stops: I'm not in the room, so I can't correct you mid-set. It costs you more time each week than an app does: the weigh-ins, the photos and the call. It costs money, and I don't publish a price list. And if you already train consistently on your own and know how to adjust your plan, a coach may not be worth it for you yet. (Here's how I'd decide.)
What the research can't tell you
I couldn't find a study that compares an online coach, a fitness app and an AI chatbot as products, over months, with people training on the plans. The studies above each test one piece: apps against no app or usual care, chatbot plans graded by experts, an app against a coach beside you. None of them can tell you which one will work for you.
So here's my read, and it's an opinion, not a finding: a plan is easy to get. All three will give you one. The question is what happens in week five, when the plan meets a bad week. An app logs it. A chatbot knows only what you type. A coach notices and changes the plan with you.
How to choose: four questions
- Do you know what to do, and do you do it? If you train consistently and mostly need a log and a nudge, an app may be all you need.
- Can you judge a plan? If you know the basics, a chatbot can save you time on a first draft. Give it detail (days, equipment, experience, goal) and check the answer. One check: the World Health Organization recommends muscle-strengthening work for all major muscle groups on 2 or more days a week for adults.10
- Do you keep starting over? If you know what to do and still stop, a better plan usually isn't the missing piece. Someone reading your numbers is. That's the case for a coach. (Why people keep starting over.)
- Is anything injured or medical? Doctor or physical therapist first, whichever of the three you pick. I'm a fitness coach, not a doctor.
It isn't either/or. You can start with an app or a draft from a chatbot, and look for a coach if the plan keeps dying in the first bad week. If you're comparing coaches, here's what to check. If you're weighing a coach against a trainer at your gym, that comparison is here.
Common questions
Is an AI personal trainer as good as a real one?
I can't tell you that, and no study I found compares them over time with people training on the plans. What I found is mostly experts grading chatbot plans on paper: general and safety-conscious, but in need of fine-tuning, less precise for an individual's goals, and better when the prompt has more detail. A person adds something a draft can't: someone who reads your numbers every week and changes the plan with you.
Can I use an AI chatbot to write my workout plan?
Yes, as a draft. Tell it your days, your equipment, your experience and your goal, and check the answer against the basics. Don't use it for anything that needs a doctor or a physical therapist; Google's own terms, for one, say not to rely on its services for professional advice.
Is a fitness app enough?
For some people, yes. If you already train consistently and mostly need a place to log your sessions, an app does that well. If you keep stopping when your week changes, the app will log the gap, but it won't ask you what happened.
Is an online coach worth it compared with an app?
It depends on what's missing. If it's the plan, an app or a chatbot may do. If it's follow-through, the coach is what you'd be paying for. Here's how I'd decide whether a coach is worth it at all.
Is there a coach page about accountability?
Yes. Accountability coaching lays out what someone to answer to looks like week to week with me.
Where to go from here
If you've tried the app and the chatbot and the plan still keeps dying, apply and tell me about your week. If an app is all you need, I'll say so. You can see how coaching with me works, step by step, or read what online fitness coaching is first.
General education only, not medical advice. If you have a medical condition, are pregnant, or take medication, check with your doctor before changing how you train or eat.
Sources
- Laranjo L, Ding D, Heleno B, et al. Do smartphone applications and activity trackers increase physical activity in adults? Systematic review, meta-analysis and metaregression. British Journal of Sports Medicine, 2021. https://pubmed.ncbi.nlm.nih.gov/33355160/ (accessed October 23, 2026).
- Cox ER, Beacroft S, Jansson AK, et al. Effects of mHealth interventions to prescribe resistance training: a systematic review and meta-analysis of randomized controlled trials. International Journal of Behavioral Nutrition and Physical Activity, 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12836956/ (accessed October 23, 2026).
- Damschroder LJ, Buis LR, McCant FA, et al. Effect of adding telephone-based brief coaching to an mHealth app (Stay Strong) for promoting physical activity among veterans: randomized controlled trial. Journal of Medical Internet Research, 2020. https://pmc.ncbi.nlm.nih.gov/articles/PMC7435619/ (accessed October 23, 2026).
- Washif JA, Pagaduan J, James C, Dergaa I, Beaven CM. Artificial intelligence in sport: Exploring the potential of using ChatGPT in resistance training prescription. Biology of Sport, 2024. https://pubmed.ncbi.nlm.nih.gov/38524820/ (accessed October 26, 2026).
- Havers T, Masur L, Isenmann E, et al. Reproducibility and quality of hypertrophy-related training plans generated by GPT-4 and Google Gemini as evaluated by coaching experts. Biology of Sport, 2025. https://pubmed.ncbi.nlm.nih.gov/40182716/ (accessed October 26, 2026).
- Dergaa I, Saad HB, El Omri A, et al. Using artificial intelligence for exercise prescription in personalised health promotion: A critical evaluation of OpenAI's GPT-4 model. Biology of Sport, 2024. https://pubmed.ncbi.nlm.nih.gov/38524814/ (accessed October 26, 2026).
- Google. Google Terms of Service (effective July 30, 2026). Google, 2026. https://policies.google.com/terms (accessed October 26, 2026).
- Harkin B, Webb TL, Chang BP, et al. Does monitoring goal progress promote goal attainment? A meta-analysis of the experimental evidence. Psychological Bulletin, 2016. https://pubmed.ncbi.nlm.nih.gov/26479070/ (accessed October 18, 2026).
- Gavanda S, Held S, Schrey S, et al. Optimizing resistance training outcomes: comparing in-person supervision, online coaching, and self-guided approaches: a randomized controlled trial. Journal of Strength and Conditioning Research, 2025. https://pubmed.ncbi.nlm.nih.gov/40728831/ (accessed October 18, 2026).
- Bull FC, Al-Ansari SS, Biddle S, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. British Journal of Sports Medicine, 2020. https://pmc.ncbi.nlm.nih.gov/articles/PMC7719906/ (accessed October 11, 2026).
