Sitch

Introduction

Sitch is an AI-powered dating app acting as a personal digital matchmaker. Instead of casual swiping, users get matched through an AI chatbot trained by a human matchmaker for deeper compatibility, targeting those serious about finding a partner and looking for a more thoughtful process.

The experience starts with in-depth onboarding, where the AI asks questions to understand a user’s values, priorities, and past dating experiences. The app chats with you like a friend to learn about you and your ideal partner. Using these answers, the AI suggests matches. If two users agree to be introduced, the AI places them in a group chat, much like a mutual friend would. The app uses a pay-per-setup model, charging users for successful introductions instead of a subscription.

Challenge

Five months after their soft launch, Sitch was gaining traction but also accumulating negative user feedback. As a developer-focused team, they lacked a structured quality assurance process, which led to bugs and usability issues slipping into the live app. Sitch partnered with QAwerk to establish an effective testing program from the ground up, aiming to transform the user experience and stabilize the product.

By the end of our partnership, Sitch wanted to achieve several key outcomes:

• Dramatically Reduce Negative Feedback: The primary goal was to improve user satisfaction by identifying and eliminating the bugs and frustrating user flows that were causing poor reviews.
• Establish a Robust QA Process: Implement a structured and efficient testing workflow that could be integrated directly with their development cycle, ensuring no new feature or fix would introduce new problems.
• Ensure Flawless Core Functionality: Guarantee that essential features, like the AI-powered onboarding, profile creation, chat, and matching, were intuitive, stable, and worked perfectly every time.
• Achieve Consistent Cross-Device Performance: Deliver a seamless and visually polished experience for all users, regardless of their iPhone model, iOS version, or network conditions.

Solution

Our approach was to embed a comprehensive and multi-layered QA strategy directly into Sitch’s development process. We focused on several types of testing, each chosen to address the unique challenges of an AI-driven, premium iOS dating app.

Functional Testing

We conducted functional testing of all critical user flows. This included validating the entire onboarding quiz, testing the profile creation and editing in the “What makes you you” section, and ensuring the account deletion process was smooth. We also rigorously tested the chat and payment screen options for new users. Our team proactively triaged issues, retested critical bug fixes, and reported new bugs we discovered along the way.

Why it matters: For an app like Sitch, core functions define the entire user journey. If a user can’t complete their profile, chat with a match, or understand the AI’s suggestions, the app fails. We needed to ensure every user path was logical, intuitive, and stress-free.

AI Testing

We performed AI testing on the AI-powered chat and onboarding quiz. Our team validated the AI’s tone and content relevance, ensuring responses were appropriate and helpful. We also ran tests to find logic flaws, such as looping or repeated questions in the quiz, to guarantee the user’s interaction with the AI was always coherent and effective.

Why it matters: Sitch’s core value proposition is its AI matchmaker. If the AI’s conversation is repetitive, irrelevant, or illogical, it breaks the user’s trust and devalues the entire “human expertise” promise. This testing was essential to ensure the AI felt intelligent and personalized.

Regression and Pre-Launch Testing

Before each release, we performed extensive regression testing on all core flows: registration, onboarding, chat, setups, and payments. For major rollouts, like the Chicago pre-launch, we executed a comprehensive checklist. This included validating market-specific copy, testing subscription pricing for the new region, verifying that referral links and analytics events were firing with the correct market identifiers, and ensuring deep links from push notifications routed users to the correct screen.

Why it matters: As Sitch’s developers fixed bugs and added new features, it was critical to ensure these changes didn’t inadvertently break existing functionality. Regression testing acts as a safety net, protecting the user experience from unintended consequences with each new app build. This became even more important when launching in new cities, where a single bug could derail the entire market expansion.

Compatibility and Permission Testing

We performed compatibility testing across various network conditions, simulating poor, unstable, and nonexistent internet connections to ensure the app remained responsive and didn’t crash. We also tested on different devices to guarantee UI elements like padding and text blocks displayed correctly on both large and small screens. Finally, we verified that crucial app permissions for the gallery, location, and notifications worked as expected.

Why it matters: Dating app users are on the move. They switch between Wi-Fi and cellular data, travel through areas with poor reception, and use a wide range of iPhone models. The app must be resilient to these changes, providing a stable experience whether the user has a perfect connection or a single bar of service.

UI Testing

Our team meticulously examined every screen to ensure a consistent and high-quality user experience. We identified and flagged issues like incorrect padding on smaller phones, awkwardly sized summary blocks, and misaligned information on profile preview screens. This focus on detail ensured the app looked and felt professional on any device.

Why it matters: The world of dating apps is quite competitive, making visual appeal and ease of use non-negotiable. A confusing or clunky interface can cause a user to abandon the app before the AI even has a chance to work its magic. The experience needs to feel polished, modern, and intuitive.

Exploratory Testing

After each release and after regression testing was complete, our team performed manual exploratory testing. We intentionally deviated from standard user paths, trying unusual combinations of actions, testing referral link edge cases, and stressing the app to uncover hidden bugs, usability gaps, and potential crash scenarios that scripted tests would miss.

Why it matters: Scripted tests confirm that functionality works as expected, but exploratory testing uncovers problems you weren’t looking for. For a high-end app, hidden bugs and usability frustrations are unacceptable.