While scaling biped.ai, an issue that came up was inbound lead management. Even though we're targeting a niche market, we get a significant inbound traffic and more users contact us for the same reason: they want to ask some technical questions, and test the device.
This is why we set up a distribution network around the world. However, for quite some time, we did all redirections by hand, which of course was painful and could involved waiting several hours, up to a day, to reply to a user and introduce them to the local distributor.
I'll present some of the "street smart" techniques I designed to help me serve leads faster & increase the traffic for our distributors.
A first technical answer that I vibe-programmed with Claude on Cursor was this:
A simple country selector with an automatic translation of the content of the website in various languages. This allowed me to display specific information with clear CTAs everywhere on the country-specific website:
This change already reduced by about 50% the amount of inbound requests that we handled directly, and when users clicked on Write an email, we are automatically added in CC.
Now on top of our very extensive FAQ, some users still had specific question and wanted to talk to someone. I looked up standard chatbot solutions, and after a long search, went for a service called Lindy that allows you to deploy GPT customer support agents that have access to your documentation in production.
The flow I implemented with this tool is the following:
This bot is live on biped's website and to date, it handles 100% of inbounds with very minimal supervision needed on top to correct errors or redirect users to new potential distributors who are not active yet for example.
This fun side-quest led to a nice time gain internally (2 to 3 hours weekly).
