Lately I have been using AI tools more and more in my day to day work, and it got me thinking about how they will change the way people buy bikes. My take is that the brands with the most to lose are not the big players, but the regional and mid-tier brands that a lot of us grew up with.
How bikes used to get sold
Let’s be honest, for a long time the typical bike buyer didn’t know much about what they were buying. Most people walking into a shop couldn’t tell you the difference between Rival and Force, or what rim width does for tire shape. They didn’t have to.
They relied on emotion and on the people around them. The shop owner recommended something. The club rode a certain brand. Maybe the bike was made thirty kilometers down the road. Trust did most of the selling.
This information gap is where a lot of regional and mid-tier brands made their living. With a good dealer network, some heritage, and a nice paint job, you could hold your price, as long as the shop believed in you. The spec sheet rarely lost a sale, because the customer rarely read it.
Enter the chatbot
For the uninitiated, large language models (LLMs) are the technology behind tools like ChatGPT, Claude, and Gemini. You type a question in plain language and get a plain language answer back.
The internet already made it possible to research a bike purchase. Forums, YouTube reviews, and comparison sites have been around for years. But it took effort, and most people didn’t bother.
Now a buyer can type “compare these three gravel bikes at €3,500, which is the best value?” and get an answer in ten seconds. The AI reads the spec sheets, lines them up, and tells them who is offering more for the money.
The problem is that these tools are great at comparing things you can measure, like groupset, weight, wheels, and price. They are not so great at valuing the stuff you can’t, like heritage, a family workshop, or the relationship with your local shop. I am already seeing this in my work with brands. Product data and spec fields now matter as much as the photos, because that is what the AI reads.
Who gets squeezed?
When buyers compare on spec and price, the advantage goes to whoever can put the most bike on the road for the least money. That is a scale game.
The big brands buy groupsets in huge volumes, spread their mold costs over tens of thousands of frames, and in many cases sell direct. A regional brand making a few thousand frames a year pays more for EVERYTHING. As I wrote in my Campagnolo article, less volume means higher costs per unit, which means lower margins, which means less money to invest in the next product.
Here is how I see it shaking out by segment:
Premium/Enthusiast: This end of the market should be fine. Nobody buys a €12,000 custom frame off a comparison table, and brands like Colnago and Pinarello have a loyal following that no chatbot will talk out of their purchase.
Entry Level: Protected by volume and price. The big brands and the new wave of Chinese brands will dominate here regardless.
The Middle: This is the danger zone. The €2,500 to €6,000 bike from a brand without scale, sold mostly on a dealer recommendation. These buyers were never fully committed to the brand, they were persuaded at the counter. Now they walk in with an AI comparison on their phone.
Is this generational?
Partly. Older riders grew up buying bikes in shops, and a lot of them will keep doing it that way. They will keep many mid-tier brands going for a while yet.
Younger riders have never known a world without a search bar. For them, asking an AI before a big purchase is just normal.
With that said, I don’t think brands can count on a slow handover. AI is being adopted across all age groups much faster than e-commerce or social media were. My dad is already asking ChatGPT about his retirement. He will ask it about his next bike too.
Keys to survive
Fix the Value: If your bike has a lower groupset tier than the competitor at the same price, the AI will point it out. Either fix the value or change the conversation.
Make it Measurable: “Handmade in Italy” is a feeling. Lifetime frame warranty, crash replacement, custom geometry, and local service are facts an AI can actually compare.
Product Data is Marketing: Clean titles, complete specs, and honest geometry charts, consistent across your website, retailers, and marketplaces. This is what gets quoted back to buyers now.
Own a Niche: The go-to answer for one discipline or type of rider will show up in AI recommendations far more often than a decent option for everything.
Partner Up: Shared sourcing, platform frames, partnerships, and even mergers are not dirty words. Going it alone without scale is going to get harder and harder.
The story behind a brand still matters, but it is no longer enough on its own. The brands that figure out how to get that story into the data will HOPEFULLY come out the other side stronger.
Like this content? Sign up for the ‘Business of Cycling’ newsletter and receive updates right in your inbox.