Hints Onboarding: A Balanced Approach to Quick Setup
Dribbble Case Study: Simplifying Complex Product Integrations via Messenger with AI Assistance
(hints.so)
Background: We were tasked with designing a product that simplifies the setup process for integrations, aiming to be more user-friendly than Zapier or IFTTT. The unique challenge was enabling users to manage these integrations through a single chat with an AI assistant, which determines which integration the user wants to activate.
Challenges:
Routing System Explanation: Our system automatically routes information, but for users desiring certainty about the information's destination, we introduced the use of a hashtag (e.g., #notion).
Simplifying Setup: The setup required users to select various parameters for each tool. We implemented backend auto-mapping using a local neural network model to aid this process. However, user authorization in the service and specific selections on updates were still needed.
Paywall Placement: We decided to introduce the paywall at the final onboarding stage, allowing users to execute several integrations to understand the product's value before guiding them to authorization and payment.
Outcomes and Learnings:
First Iteration of Onboarding: This initial approach was quickly revised due to several issues:
Messenger Authorization Disconnect: Users, once directed to the messenger for authorization, tended not to return, leading us to integrate the remaining onboarding steps directly within the connected messenger in the next version.
Delayed Authorization Flaw: Moving the authorization to a later stage proved to be a mistake. We lost the ability to send follow-ups, impacting our understanding of user activities during onboarding.
Complex Steps: Despite efforts to simplify, the steps remained too complex. We had to eliminate many of them, including the explanation of our newly coined term "flow".