Both AI and no-code are technologies that digital product teams can make use of to drastically decrease costs and boost the speed of exploring and testing new ideas. But is it feasible to combine them and how? Are there already commercial no-code platforms that have AI integrated? And how big is the combined benefit? We’ve taken a pragmatic approach to answer these questions and challenged ourselves to create an app called Bobcatter, a replica of X.com – relying solely on these cutting-edge technologies, only in 3 weeks and of course, without writing a single line of code.
An imminent paradigm shift in software development is that software in the future won’t be created through coding or even through drag-n-drop – but through conversation. You’ll be able to describe your app in natural language to the no-code platform, it’ll then generate a base no-code app as a starting point and afterward, you and the platform will be able to iterate on it in a conversational way. This is already happening, as the most advanced platforms today allow anyone to generate a tailored website or a step-by-step guide to follow to develop apps – both based on text prompts and in seconds.
The term “AI-assisted no-code”, of course, can be interpreted in a broader sense as well: you can use AI to support the development process not only during building the software itself in the no-code platform but also both before and after that. For instance, there are AI tools to help you create the visual assets and the wireframes that can be provided as inputs to no-code, while others are for aiding you with evaluating the output, namely producing the test cases and generating the corresponding test code.
We approached “AI-assisted no-code” in this broader sense when building Bobcatter. We experimented with several AI tools like Canva and Uizard to create Bobcatter’s logo and its interactive prototypes, whereas we used TestCraft and Testim to check the quality of the no-code outcome. With regards to the no-code platform, after much deliberation, we decided on Bubble.io because it’s one of the most well-known in the no-code ocean, has extensive documentation along with a large user community, and is an all-in-one solution with a promise of rapid development for both the front-end and back-end.
With Bubble.io's AI tool, we were able to generate a base UI that was remarkably close to what we had envisioned for Bobcatter in seconds, taking a fraction of the time that coding from scratch would have needed. Essential functionalities like user authentication and database management, which would typically require significant backend work, were essentially built into the platform, and thus we could implement tasks that otherwise take days of coding, in hours if not minutes. In addition, using the plugin ecosystem of the platform, we were also able to deploy advanced features like Google SSO relatively quickly through pre-built plugins.
On the flip side, we encountered some potential limitations of the approach as well. The inability to add custom code or easily migrate to another platform felt like a possible straitjacket. The difficulty with handling complex, custom features and the lack of comprehensive version control are also likely long-term concerns. Furthermore, the potential need to rebuild the app from scratch if we needed to transition to a traditionally coded version is a significant risk to consider, too. Finally, as many of the AI-powered testing tools seem to be designed with traditional, code-based applications in mind, making them play nice with no-code requires some creative thinking and workarounds.
The speed at which we could iterate on different versions and feature sets of Bobcatter thanks to AI-assisted no-code was remarkable and this quick turnaround could be invaluable for small teams looking to validate their market quickly, launching prototypes and MVPs to gauge interest before making significant investments.
However, we also recognized that this approach is not a one-size-fits-all solution, and for applications requiring high levels of customization or scalability, traditional coding might still be the better option. We see a continued need for skilled developers, and indeed, as these technologies continue to evolve, we believe the most successful developers will be those who can navigate both worlds, making strategic decisions about which approach best serves a project's needs at different stages of its lifecycle, be it an AI-assisted no-code platform, traditional coding, or a hybrid approach.
To immerse yourself in our adventurous Bobcatter journey, read our diary-style white paper “Journey into AI-Assisted No-Code Development: Creating Bobcatter”, written in a language and structure that can be understood by non-coder human creatures as well.
Want to experience Bobcatter firsthand? You can leave us a Purr right here.🐾
Databutton (2023): Ditching the drag-n-drop: The No-Code Revolution No One Saw Coming
Comments