AI route explanations: let Cyclemate tell you why it picked your route

Ever looked at a Cyclemate route and wondered why it sent you one way and not the obvious way? You're not the only one. It's one of the most common things riders ask us, and it's a fair question. So we're rolling out AI route explanations: a plain-language line on every route that tells you where it comes from.

A lot goes into every route

Behind every set of directions is a routing model we build for each city. It prioritises protected cycleways, steers around dangerous junctions and collision hotspots, prefers well-lit streets after dark, avoids uncomfortable hills, and factors in reports from riders like you. Every road is graded (quiet, busy, or very busy) and we stitch together the route that keeps you on the calm ones.

That's a lot of judgement calls happening in a fraction of a second, and until now they all happened quietly under the hood. You got a good route, but you couldn't see the reasoning.

Now it shows its working

Tap a route and you'll see a short summary of the key difference: the safer road it takes and the busy one it steers around. Something like "Takes the C5 cycleway and avoids busy Haggerston High Street." We use AI to turn what the routing model already knows into a sentence you can read at a glance.

It's not there to drown you in detail. It's there so you can trust the route in front of you, and when we show you a few options, so you can see at a glance what makes each one different and pick the one that suits your ride.

This is the first step in a bigger push on transparency: showing our working, opening up the data behind our directions, and giving you the context to make your own call. It pairs nicely with the Explore map we launched recently, which lays out the datasets underneath every route.

Route explanations are rolling out now, starting in London. As always, we'd love to know what you think, so reply to this email or DM us.

Cheers, The Cyclemate team