how Flow (DapperLabs) automates answering +1k developer questions every month


~2 support engineers saved by having implemented per month (assuming avg. 15mins per question, ~1100 questions per month, 160 working hrs per month)

~140 days of end-user waiting time saved per month from instant answers (assuming avg. human response delay of 3hrs, ~1100 questions per month)

98% accuracy of questions answered by based on user-rated feedback



Rapid developer community growth puts a burden on support

Flow, a fast, decentralized, and developer-friendly blockchain, designed as the foundation for a new generation of games, apps, and the digital assets that power them, is one of the most popular web3 projects. However, the growing developer ecosystem led to an increasing number of technical questions, placing a larger support burden on the core engineering team.



A GPT-4 powered chat widget trained on Flow knowledge sources

Knowledge Sources worked with the Flow team to identify relevant knowledge sources to feed into a chatbot including their docs site, SDKs, and more than 90 transcribed YouTube videos.

Deployment Method

The chatbot was then deployed directly on the Flow docs as a chat widget as well as in their Discord community. With the addition of the docs widget, usage grew approximately 1000%, going from answering around 20 questions a week to more than 350 questions a week as more users became aware of the widget.

Try directly on the Flow docs site: [Flow Docs](

Try directly in the public Flow Discord community: [Flow Discord](


THE RESULTS handles ~1.1k questions per month's chatbot is a game-changer for the Flow team. The chatbot has currently answered over 1100 questions with an estimated accuracy of 98%, significantly reducing the support hours needed. This allowed Flow to reduce support hours by +300hrs, freeing up the team's time to focus on more strategic tasks, such as developing new features and improving the coin. The Flow community is thrilled with's solution, enabling them to get instant and accurate responses, improving their overall user experience.


  • See core team member Dmitrii tweet about the new integration on the flow docs site

  • Hear more about Flow in our podcast:


Flow Developer Docs


California, USA


Web3 Framework

Deployment Method

Discord Bot

Docs Widget