How Does TrainerRoad Pick Your Workouts?
TrainerRoad picks each workout by running hundreds of simulations on your own training history. It predicts how every workout it could give you would go, based on the rides you have done, then puts the one that best fits your goals and your current fitness on your calendar. That happens whether you follow a plan or just ask for a ride on the day.
If you are deciding whether to trust that pick, the fair questions are the ones athletes ask us all the time. Is it just a template? What does it actually look at? And how do you know the workout it chose is the right one for you today? This post answers each of those in turn.
Is TrainerRoad Just a Templated Plan?
No. A TrainerRoad plan sets your goals and your schedule, but the workout on each day is picked for you after TrainerRoad runs hundreds of simulations across all possible workouts it could give you, all in an effort to find the workout that most closely fits your goals and current fitness level. So the workouts you get are not copied from a fixed file. Two athletes on the same plan, even with the same event and the same hours, would likely get different workouts because they are different athletes.
That is the difference between a template and a plan built around you. A template hands every athlete the same Thursday. TrainerRoad looks at your Tuesday first, and your last few weeks before that, runs simulations to have confidence rather than guess, and then decides what Thursday should be. If you took an extra rest day, or did a hard group ride that was not on the calendar, Thursday changes with it. The way TrainerRoad builds plans and the way it picks a single AI Workout are the same process.
What Is TrainerRoad AI?
TrainerRoad AI is the set of models TrainerRoad trains on its own dataset of millions of rides from real athletes, from beginners to WorldTour professionals, so it knows what kinds of workouts make cyclists faster in every scenario. It pairs that knowledge with your individual training history to learn how you respond to training and what balance of volume and intensity works best for you. It gets to know you well enough as a rider to predict how you will respond to a given workout, and it uses that to run simulations across every workout you could get and find the right one each day.
What Does TrainerRoad Look At When It Picks a Workout?
The simulations start with every ride you have done. Indoor workouts, outside rides, and Zwift races all count, and TrainerRoad reads the power, the heart rate, and how you said the ride felt. On top of that it uses how fatigued you are right now, your schedule, and what you are training for.
From there it simulates every workout it could give you for that day. For each one it predicts how hard it would feel for you, if it is giving you progressive overload in the areas that are most relevant to your goals, how much training stress it would add, and what it would do to the workouts that come after it. Then it picks the workout that gives you the best chance of making the progress you want. The AI Workouts post goes deeper on that selection step if you want it.
How Does TrainerRoad Pick Workouts When You Follow a Plan?
When you follow a plan, you tell TrainerRoad your goals, your event, how many days you can ride, and how long you want to train each day. It uses all of that to build a custom training plan for you with proper periodization so you peak at the right time and are well-prepared for your goal event. It then uses that periodization framework and everything you told it about your schedule, and runs simulations for each training day on your calendar to know what workout is the right one for today.
You can see this in your calendar. The next four weeks show the workouts TrainerRoad currently expects to give you, and that view is the AI Training Simulation. It is the best current prediction, and it moves every time you ride or change something. Set your plan up with Plan Builder and the rest is picked for you. If you are not racing, Plan Builder works for non-racers too.
How Does TrainerRoad Pick Workouts Without a Plan?
Without a plan, you tell TrainerRoad how long you have and how hard you want to ride, and it picks the workout for you with the same simulations. You choose a zone, like Endurance, Sweet Spot, or Threshold, and a duration, and TrainerRoad runs the simulations across every workout in that category and schedules the best one.
That is an AI Workout: a single workout TrainerRoad picks for you from a zone and a duration, with no plan behind it. It is the right tool for a week with no structure, a few weeks between plans, or a day you want to add on top of your plan. You still get a workout that was picked from your history, not from a list of favorites. If the workout it picked does not fit your day, you can ask for an alternate that is shorter, longer, easier, or harder, and the swap re-simulates the days after it. It won’t be part of a larger periodized plan or prepare you for a specific goal, but it will give you the right workout on the day.
With a plan
Picked to build toward your goal
- You tell it
- Your event, your days, your hours
- It picks
- Each day’s workout in your plan
- What shapes the pick
- Your rides and the plan’s periodization
Without a plan
Picked for where you are today
- You tell it
- A zone and a duration
- It picks
- The right workout for today
- What shapes the pick
- Your rides, with no periodization
What Changes After You Finish a Ride?
Every ride you finish feeds back into TrainerRoad’s AI, informing and adjusting your upcoming workouts. If the ride went the way TrainerRoad expected, you will not see much move. If it took more out of you than planned, the next workout changes to protect your recovery and keep your fitness moving forward. If it felt easier than expected, the next workout can step up.
The clearest example is a hard ride that was not on the calendar. Say you bury yourself in a group ride on Monday. By the time you open the app, Tuesday’s workout has already been replaced with a recovery day. That is Fatigue Detection, the part of TrainerRoad AI that spots fatigue before it costs you a workout. It looks at the load you just added and flags the recovery day before you have noticed your legs are gone. The same thing happens in the other direction. Skip a workout and TrainerRoad does not pile it onto the next day. It re-simulates from where you actually are.
What Does the Post-Workout Survey Do?
The post-workout survey gives TrainerRoad information your power file cannot show, which makes the next pick more specific to you. After a workout, TrainerRoad asks one question: how did this effort feel? Your answer, from Easy to Maximum Effort, tells it whether the ride landed the way the simulations predicted.
Two riders can finish the same workout with the same power and have very different days. One was cruising. The other was hanging on from the second interval. TrainerRoad can already see some of that difference in how each rider held the power and what their heart rate did, and the post-workout survey adds another layer of insight on top. Rate a workout Very Hard that it predicted would be Hard, and the next pick accounts for that. On some workouts you will get a second question asking why, which adds even more detail. Answering takes a few seconds, and it is the single easiest thing you can do to make your workouts fit you better.
How Do You Know the Workout Is Right Before You Start?
Every workout on your calendar carries a Predicted Workout Difficulty, from Easy to Maximal, so you know what you are walking into before you clip in. That label comes from the same simulations that picked the workout. TrainerRoad has already estimated how the workout is likely to feel for you, given your recent training and your recovery, and it shows you that estimate in plain language.
Open the workout and you can see the odds of each outcome, which is useful when you are deciding whether today is the day for it. A workout predicted to feel Very Hard on a Tuesday after a race weekend is TrainerRoad being honest with you, and that is usually your cue to take the recovery day it is offering instead. The predicted difficulty post explains how to read those labels.
Frequently Asked Questions
What is TrainerRoad AI?
TrainerRoad AI is the set of models TrainerRoad trains on millions of real rides, from beginners to WorldTour professionals, to learn what makes cyclists faster. It pairs that with your own training history to predict how you will respond to any workout, then runs hundreds of simulations to pick the right one for you each day.
Does TrainerRoad use AI to pick workouts?
Yes. TrainerRoad AI runs hundreds of simulations on your own training history to predict how each possible workout would go for you, then picks the one with the best predicted outcome for your goals and your current fitness. It does this for every workout in a plan and for every AI Workout you add without one.
Do I need a training plan to use TrainerRoad?
No. Without a plan, you can add an AI Workout by choosing a zone and a duration, and TrainerRoad picks the specific workout from your training history. A plan adds your goals and your schedule, so TrainerRoad can pick every workout for weeks ahead instead of one day at a time. Either way, the right workout shows up in Zwift or Rouvy for you automatically every day.
Does TrainerRoad look at my outside rides?
Yes. Outside rides, group rides, and Zwift races all go into the simulations alongside your indoor workouts, as long as they sync to your TrainerRoad calendar. A hard ride that was not on the plan changes the workouts that follow it, the same way a planned workout would.
What happens if I skip a workout?
TrainerRoad re-simulates the next four weeks from where you actually are. It does not stack the missed workout onto the next day. If the skipped day means you are fresher than expected, the next workout can get harder. If you skipped because you were cooked, the post-workout survey on your last ride helps it see that.
Why did TrainerRoad give me a recovery day I did not plan?
Because Fatigue Detection predicted that a hard workout on that day would not be productive. It looks at the load from your recent rides, including ones that were not on the calendar, and swaps in a recovery day before you have noticed you need one. You can still ride what you want, but the recovery day is the pick with the best predicted outcome.
Can I change the workout TrainerRoad picked?
Yes. Open the workout and choose an alternate that is shorter, longer, easier, or harder, and TrainerRoad re-simulates the days after it. You can also add an AI Workout in a different zone or move workouts around your week, and the calendar updates to match.
Does the post-workout survey actually change anything?
Yes. The survey tells TrainerRoad how a workout felt, which the power data alone cannot show. Rate a workout harder or easier than it predicted and the next picks account for that, so the more consistently you answer, the more specific your workouts get.
If you have ever stared at a workout library wondering which one to do, this is the part TrainerRoad takes off your plate. It reads your rides, runs the simulations, and puts the workout with the best chance of making you faster on your calendar. Your job is to ride it, tell it how it felt, and let the next one get better.
TrainerRoad finds the right workout for every ride and adjusts to you in real time. Try it risk-free for 30 days.
Start Getting Faster
