In many fields, AI assistants in the form of chatbots are simply inconvenient. For example, for designers, video creators, or architects, platforms with a completely different interface are needed, within which special AI machines should be embedded. Here’s an example of such a platform for designers. And it prompts interesting thoughts for applying this approach in other areas.
Project Essence
Pimento calls its platform a “creative AI companion”.
Its target audience includes creative agencies, design and marketing teams in companies, as well as freelance designers.
The essence is that the creation of an advertising campaign, website, branding for a new product, and other creative materials typically starts with a very general picture – how it should all look in principle, what mood to convey, what types of images and what colors should be used. The detailing of specific design elements begins only after such a general picture is discussed and approved.
This “general picture” is communicated and discussed not in words. For this, designers use a “moodboard” – which can be loosely translated as a “mood board”. A moodboard is a real or digital board to which images, color samples, fonts, and other design elements are attached – which, taken together, convey a certain mood and set the general direction for further creative work.
Previously, platforms like Pinterest were often used for this, where one can create digital thematic boards and attach internet-found images on a theme.
Pimento is a more convenient tool that allows designers to create, discuss, and share such “mood boards” with clients and the team.
The first charm of the platform is that designers do not need to search the internet for images and design elements that exactly fit the desired mood. They can start with a few images somewhat matching the desired mood and then ask the platform’s AI machine to generate more suitable images with the necessary changes.
Understandably, even a very smart AI machine won’t generate something exactly suitable on the first try. It doesn’t know what is really hidden in the designer’s mind, and even the designer himself doesn’t know for sure yet 😉
Therefore, the platform now allows you to “play” with the set of images created in the first stage – for example, ask the AI machine to change the color, change details, clarify the mood of the image, or even create a new image from several more or less suitable ones.
Additionally, the AI machine can be asked to generate examples of colors and fonts suitable for the chosen mood.
Thus, through a series of refinements, designers gradually form a set of images, colors, and fonts that most closely match the desired design image.
From the most successful elements, separate boards can be created, which can be shared with team members and clients in one click – and they, in turn, can leave their comments on the board, in the form of texts as well as new images.
The minimal version of the platform with two boards and 100 generated images per board can be used for free. 3 boards and 200 images per board cost $25 per month, 20 boards and 500 images per board – $129 per month. There is also an unlimited option – but for its prices, you need to contact the startup.
Pimento was created in October last year, and then it received small pre-seed investments. And now it has raised another 3 million euros in new investments.
What’s Interesting
Yesterday I wrote a review of the startup AutogenAI, which helps create compelling tender applications, commercial offers, and other sales documents using a specialized AI machine.
An important feature of the platform’s AI machine, I called it, is that the final result is created on it “by the method of successive approximations”.
First, the AI machine generates a list of main arguments, which can be edited – throwing something out, adding something. Then for each top-level argument, it generates a list of more detailed theses, which can also be edited. And only then, based on the created “skeleton”, a final coherent text in human language is generated.
Interestingly, today’s Pimento platform uses a similar concept! First, the designer sets the starting points for creating a “mood board” and asks the AI machine to generate a set of images. Then he chooses the more or less suitable ones and again asks the AI machine to do something with them – so that step by step he reaches the desired result, which he can already share with colleagues and clients.
So it turns out that the method of successive approximations is a working concept, which even finds its application in such an unexpected area.
The concept of “mood boards” and the use of AI as a “creative companion” is also used by the startup MagicBrief. However, they chose a narrower field for themselves – creating advertising videos, and the sources of inspiration here are other people’s videos, which can be automatically dissected “to the bone”, and then reassembled under oneself. This startup raised $2 million in its first round of investments.
An important feature of graphic designers and videographers (video creators) is that they generally think in images. Therefore, to support their creative process, they need AI assistants who also think in images – unlike the textual ChatGPT, for example. It is this fact that is exploited by today’s Pimento and MagicBrief.
Interestingly, such imaginative thinking is also characteristic of architects and builders. This is noted by the startup Planera, which I wrote about in May. Its creators claim that traditional project management platforms, showing projects in the form of huge tables, do not suit architects and builders – precisely because they think in pictures. And therefore, they need special project management platforms, in which projects will be presented in the form of images, not tables. They made such a platform – and raised $5.4 million in the first round of investments.
Where to Run
First of all, I will again draw attention to the method of successive approximations when using AI – when we do not try to achieve the perfect result from AI by pressing one magic button, but go to it through a series of steps.
This way, in many cases, we can achieve the result we need much faster. Especially when we ourselves do not know exactly what we want to get in the end 😉 In fact, this is a natural part of any creative process – when there is a vague general picture in the head, and we gradually work it out. Sometimes we even sharply veer to the side because some new thought came to mind while working out.
And this applies not only to design or video shooting. One of the famous writers said that at the very beginning of writing a book, he has only the characters of the heroes and their first conflict in his head – and then they start to live independently, and the author just records them.
Incidentally, in a similar state of “flow” I write reviews for Fastfounder – immersing myself in the topic as I write the review. As we see in the example of AutogenAI, it is not only possible but also effective to create commercial offers, tender applications, and even startup presentations in a similar way 😉
So, it seems that the topic of successive approximations with AI in the creative process is still not sufficiently revealed. For example, the vast majority of AI text generators offer to create the perfect text with one button press – which is fine for student essays, SEO texts, and other dross, but not for something really important.
The second important point is that the textual interface of communication with AI assistants is not suitable for every area. At a minimum, designers, videographers, and architects need to communicate with AI assistants in pictures – which means completely different platforms are needed, not just ordinary AI chatbots. An example of such a platform is today’s Pimento.
If you think about it, few people create directly in words and formulations. For example, how does a programmer create? Not with specific lines of code – I can say that as a programmer 😉 Writing a program starts with some vague picture of the general algorithm, which gradually manifests itself as the code is written. I wonder what an AI companion for “creating” programs, not for stupidly coding functions according to a given input and output, might look like?
In general, the direction of movement can be formulated as the creation of AI companions for creative activities.
Understanding “creativity” in a fairly broad sense – design, creating videos, writing books and articles, creating architectural projects, programming… and even creating ideas and business models for startups, finally 😉
So – what creative activity and how would you like to automate with the help of a corresponding AI companion?
About the company
Pimento
Website: pimento.design
Last round: €3M, 12/07/2023
Total investments: €3M+, rounds: 2