TrainAsONE is an AI based training plan for runners. It takes data from runs you’ve completed, using Strava amongst others, and schedules future runs based on a goal – often a race – set by the user. Each plan period starts off with three benchmark runs, often one per week, surrounded by several “economy” runs. Economy generally meaning easy.

Perceived effort run

The first of these benchmark runs is a perceived effort run. The AI system asks you to complete a run whereby you spend the first 5 minutes or so at a very easy pace, followed by a set period at your natural easy pace. It would appear that this allows the system to understand your natural running pace and allows it to calculate the pace that you should run at for your economy runs.

3.2km Assessment

The second type of run is the 3.2km/2 mile assessment. This essentially has you warm up, take in some easy miles and then push as hard as you can for 2 miles, followed by creeping up the pace from walking back to easy. I believe this benchmarking session allows the system to calculate your natural interval pace.

6 Minute assessment

The final benchmarking run that TrainAsONE (TAO) requests that you complete is a 6 minute assessment. It is similar to the 2 mile assessment in that it includes a warm up and some easy miles before an all out 6 minute time trial. One would imagine that the result of this session allows the AI system to calculate your repetition pace.

The plan

Following on from these three assessment runs, the web based software will provide you with a training plan leading up to your primary goal. I say primary as premium users can set multiple goals, with one primary, whereas free (yes, free!) users can only set the one. Standard users can see the next 14 days of their built plan, whereas premium users can view their calendar 42 days in advance.

The plan will consist of different types of runs, including economy, threshold, intervals and repetitions. Each run will include a warm up of around 5 minutes, followed by some easy running and then the session. As previously mentioned, economy runs are what would be considered as easy. Threshold runs tend to consist of intervals of 6-12 minutes, with 2 minutes of rest in between. Intervals are almost always 3 minutes long, with 3 minutes rest and repetitions are 30 seconds long with around a minutes rest. The best part about the scheduled runs is that it provides a pace target for each of the steps. See the below example for an upcoming economy run. I should point out, as you can see in the image, that the workouts can be downloaded from the system and uploaded to your Garmin watch as a workout, meaning you can follow the steps automatically from your wrist, rather than having to remember them whilst you’re out.

Setting a goal

When setting your goal, you select a distance and a date, most commonly a race. You then set your commitments. This allows you to set how many days you would like to run per week, on what days you are available to run and for how long on each of those days. The difference here between free and paid is that with the premium version you can set this to be different each week. With the free version, you set it up as a weekly occurrence, which can be changed, but then changes it for subsequent weeks. Less flexibility. You also set your weight, resting HR and maximum HR, which allows the AI to adjust your plan and paces accordingly.

Another important point to make about the plan, is that it takes into account weekly training load. In using this, assuming it is followed by the user, it does not allow for overtraining. Each week is a gradual build on the previous week, therefore reducing the risk of injury.

My experience

If I said that I had been using TAO for 2 years now, I would be half telling the truth. I signed up to the Beta program in around Q3/Q4 2016. I followed the suggested plan fairly religiously for around 6-8 months and did see results. This included around a 2 minute improvement in my 5km time and an 11 minute improvement in my half marathon time, although this also coincided with joining the Rugby Runners running club.

The reason I stopped using the software was due to repetition. I understand the need for benchmarking, and for measuring improvement, but completing the three assessment runs every 6-8 weeks became tiresome. There’s also the sessions. What I’ve mentioned above is strictly adhered to, meaning EVERY interval session includes 3 minute reps, and EVERY repetition session includes 30 second reps. There’s no variety. No 400m/800m/1km reps, no pyramid sessions and no varying rest periods. It’s all too repetitive. So whilst it did provide results, it got a bit boring in the end.

That being said, that was 2 years ago, and during the Beta stage of testing. The system is now in full flow, and with a paid subscription service too. With that in mind, I’ve decided to give it another go, as I still have my original account, and have since completed both the perceived effort and 2 mile assessment runs, with the 6 minute one to follow in the coming week.

I’m not expecting big results, and I’m unsure at this stage how long I will continue it for. If the sessions are the same as before, I imagine it won’t be for very long. The difference now, in comparison to when I first signed up to the Beta, is that I know what it takes to become a faster runner. I know how to plan my own threshold and interval sessions and I know how to avoid overtraining. Therefore I don’t NEED such a ‘training partner.’ We will see how it goes…

If you wish to explore TrainAsONE yourself, please follow the link below to their homepage. If you’re relatively new to speed work, or if you’re in a bit of a rut and wish to get faster but are not sure how, then it’s a great piece of software to give you guidance. If you’re a more seasoned runner, with experience in speed work and you know how to set up your own sessions, then I’m not sure what more it could offer you.

Categories: Blog


Andy Rich · July 1, 2019 at 13:33

I’ve been using TAO for about a month. As an experienced runner (3:12 marathon), who’s got out of the habit for the last 3 years, its provided me the motivation and kick up the arse I needed! I often used to overtrain and burnout, I’m hoping TAO will stop me doing that by intelligently increasing my training load. Plus I think its such a cool concept with AI & big data behind the scenes rather than blindly following a generic plan. Time will tell. 400/800 reps are just done because of the track length. I think the plan is quite varied so maybe its improved since you last used it.

    Paul Price · July 2, 2019 at 20:40

    I must admit, when I started using TAO, for about 6-9 months I did see some improvements in my race times, but then it all got a bit repetitive and stale almost. As you quite rightly said, it may have improved since then. I do agree that something like TAO offers more for YOU than a generic training plan. I hope it goes well for you. If you get chance, do check back and let me know how you get on, I would be interested in your long term experience

Johan · October 22, 2019 at 07:33

I’ve been using the paid version of TrainAsOne for about 4 months now. Before that I used to analyze my own data and compile my own training plan. I did 3 marathons “on my own”, but was disappointed by my own results (PB 4:25).

I am super excited about the TrainAsOne concept. It obviously depends on your personality, but I prefer AI over a human coach. When I cannot run on a weekend due to other commitments, it patiently adjusts my training to work around it. Personally I find enough variation to keep it interesting. I also find that I train less than I would have on my own, which keeps me from training to the point of fatigue.

I’ve got a race coming up in three weeks. First one on TrainAsOne. I am very curious and excited to see how it goes.

    Paul Price · October 22, 2019 at 20:19

    Sounds like it has definitely worked for you Johan, and others too. Maybe I need to give it another chance at some point. All the best for your upcoming race.

Leave a Reply

Your email address will not be published. Required fields are marked *