Using AI To Choose Your Running Shoe Needs

As a shoe fitter guru who has been around for quite a while (along with my degree in Kinesiology), never did I think there would be a day that AI would make my justification and choices of shoes invalid. What a wild world we are in! 

Firstly, this is not a hack against AI. It is a great place to start, and a great resource where you get to ask multiple questions that can start to filter the many running shoes available to us. Especially if you don't know where to begin and don't already have a relationship with your local running store. But there are limitations to AI and what it has learned so far. 

We recently had a customer come into our store looking for running shoes to fit their new orthotics. They used AI prior to stopping in and one of the answers it generated stood out: stay away from New Balance running shoes. While I agree with some of the answers it generated, I disagree with this point. For years, New Balance has some of the most recommended shoes on the market for orthotics, and their top seller within the running industry (the 880) is a staple for people wearing orthotics. This is because it has generous depth, good stability and a relatively neutral surface to support inserts. While I don't think all New Balance shoes are appropriate with orthotics, AI grouped the whole brand into their answer. 

AI is limited to the information available to it.

AI will often only give you limited answers instead of pages of choices. If you use a search engine to find products, it will give you millions of options which you can scan and figure which might work the best for you. AI gives you limited answers based on the information it has access to on the web. AI doesn't have access to scientific papers, and depending on the program you use, it might not give you the data sources that it pulled from.

I often find it amazing how all of us can feel something different in a shoe. The reasons I love a specific shoe may be the opposite of what someone else prefers, given different running mechanics, body types and paces, we may still run in the same shoes. Now, throw this information into a product review (which AI has access to) about your own personal pro's and con's and this could sway the answer that AI gives you. 

AI will give you answers based on what it deems to be reputable sources on the internet and filter these down accordingly. While I cannot ask AI who it deems reputable (please make this a feature in the future), I can sure ask a human what their qualifications are, what education they have completed, and what their personal fit philosophy is (yes, this is a thing), etc. 

I asked AI for recommendations based on my foot and my preferences, and the answers I got were not even close to the shoes that I run in, nor close to the recommendations that I would give to someone who came into the store. 

AI is limited to the information you give it. 

My preference criteria for shoes is this: neutral, stable cushion on the firmer side and usually in the higher cushion category. I'm usually running at max 20km/week and around 6:45/km. The three shoe options it gave me weren't even close to the brands that I find comfortable to run in. 

When a customer comes into a running store, you might give us (the shoe fitters) the same criteria that I gave AI and then let the shoe fitter loose to find shoes. More often than not, a shoe fitter will ask more questions to help them filter down the massive catalogue of shoes available: what shoes have you ran in the past, have you recently been fitted for shoes, what did or didn't you like about your last shoes, etc. AI didn't prompt for more information until later on in the chat where it assumed that my "firm" cushion isn't the same definition that it is using (points to AI for this). But even after clarification, it still came back with the same 3 pairs from the original search that I told it I do not like. At this point, as someone who is looking for advice, I would just be frustrated. As someone who was almost prompting AI to guess my shoes, I just started laughing. What a wild world we live in. 

While I might not be in danger of losing my job to AI, it raises the question of what or who you would deem a reputable source and then be able to trust the answers on your search for new shoes. I hope our store is one of them!!! 

My Key Points for Running Shoes:

Get fitted from a local running store. Most staff at running stores have advanced education in kinesiology or biomechanics or have undergone training by someone and have learned significant biomechanics.  Have someone watch you walk / do a gait analysis. This will allow the shoe fitter to see what your feet are or aren't doing. Get this done every few years and especially if anything has changed. Injuries, pregnancy, weight loss or gain can all change your feet.  

Try a variety of shoes based on your gait analysis. When you are searching for shoes, the best success for finding a shoe that you love is having that instant comfort in them and the shoe feeling natural on your foot. There is no break in with running shoes - if something isn't working, it's not going to get better. 

Test out the shoes where you can and make sure nothing feels off or weird. Most stores have a return policy based on clean and re-sellable and allows you to try on a treadmill or track or around your house.  

I'm sure all of this will change with the learning trajectory of AI but for now, I find asking for more advanced questions leads to unreliable answers. What a wild time we live in. 

Lauren Milburn is the general manager for Strides Running Store. She holds a B.Sc in Kinesiology, has worked in the running shoe industry since university (aka a really long time) and is a shoe nerd. She loves the weird shoes fits and will have the patience and/or persistence to find that right shoe for you!