The missing key features for designers to incorporate AI into their everyday work


3 Problems with Current AI Tools: Conclusion from 30 Interviews

After speaking with 30 designers from various disciplines, including industrial, fashion, graphic, and interior design, we have identified three main issues that designers face when trying to incorporate AI into their daily routines.

Problem 1: User Interface

When working with AI, it is common to generate multiple variations due to the unpredictable nature of the results. This is especially true for designers during the ideation and iteration process.

However, there is a major problem:

🙄 The user interface of current AI tools is not very user-friendly.

This is because the interface is not designed to accommodate an iterative process.

With current AI tools, you need to download both the prompt and the image results and transfer them to another software for better comparison, editing, and further use. Finding a specific image or prompt through the iteration process can be a tedious task, requiring you to scroll through irrelevant content.

In other words,

  1. Navigating the process can be challenging, and it can also be difficult to see the bigger picture during the iteration process with AI.
  2. You will always need to download and upload the results to another application.

As designers, we used to display all sketches and ideas on a wall for discussion, review, and feedback. Similarly, to incorporate AI into our work, we need a large enough space where we can see everything, organize ideas and graphics, experiment with different prompts, generate more graphics, and develop further—all in one place.

Answer: AI on an infinite canvas + database

An infinite canvas allows us to put everything in the same place, and a database helps us organize results and iterate on different prompts.

That’s why we’ve built AI features into the whiteboard as one of our first experiments.

The results were surprisingly good:

  1. You’re given a lot of freedom to modify prompts by combining different words,
  2. You will keep generating images till you are happy, and all images and prompts are seen in one place,
  3. When using images as the input, it is very easy to match it to the right prompt.

Problem 2: Prompt Engineering

As we all know by now, prompt engineering is a big part of working with generative AI tools. Although it may seem novel and even fun for some people, it is still “engineering”.

The vast majority of potential users of generative AI tools lack the necessary understanding and knowledge to engineer the prompt to do what they intended.

We will see a wave of new approaches to simplify the process as more applications become vertical. Vertical applications reduces the necessity to have users tailor the prompt towards specific uses and may develop into varies ways to interact with users.

Prompt engineering will still be there for the most explorative part of generative AI, but as AI evolves into a productivity tool in daily workflows, less people will need to worry about it.

Problem 3: An All-In-One AI Workspace

As more AI applications tailored to specific industries and design scenarios emerge, designers are finding themselves dealing with numerous different tools.

👀 It is essential to be able to use different tools together in one place, and for different AI systems to interact and communicate with each other.

For example, imagine feeding ChatGPT two images – one featuring what we already possess and another displaying our desired aesthetic. It can provide suggestions of prompts on how to best achieve this outcome. Alternatively, you could employ ChatGPT for advice on specific AI tools and processes that are most suitable for attaining your desired results.

With this kind of interconnectivity between AI, the possibilities of creation become endless!

AI is no longer just “one-and-done” where you offer an input, the machine spits out an output, and you can keep it or throw it away and try again. The whole workflow with AI can now become much more iterative, where you can feed outputs from one AI app to another, leveraging various AI applications to finesse and generate better variations.

In addition, as more AI tools are involved in our workflows, and as the iterative process becomes more complicated and cumbersome (e.g. experimenting with different keywords, tools, and workflows), it would be great if these repetitive jobs could be automated into streamlined workflows. It would be even better if the entire team could collaborate on AI-driven projects.

That’s what Fabrie hopes to achieve.

As an all-in-one workspace, Fabrie can help designers employ AI, automate AI workflows, and collaborate on AI-assisted design projects.