Discovering Visual Sparks through AI: Tools to Accelerate Design Workflows

At Orion, we’re exploring how AI tools are transforming the way design teams and organizations operate. Discover how these tools are applied in practice, with real-world insights from our award-winning Experience Design team.

AI is making its way into every corner of the enterprise, from powering chatbots to conducting cybersecurity to writing copy. With the plethora of tools available, how might some of these fit into the larger arsenal of creative teams, particularly a product designer’s project workflow? Let’s take a look at how a few select AI tools can accelerate early-stage product design work: exploratory processes for typography, color palettes, and logo design.

Typography

To serif or not to serif? When building a design system, eventually typefaces need to be selected as an extension of a company’s branding or implemented across an entire application. Naturally, header styles need to be considered alongside paragraph styles, and this is where a tool like FontJoy can help.

Through deep learning, FontJoy generates different font combinations intended to create a pleasing contrast among two header styles and one paragraph style. Users can adjust sliders to increase the visual disparity among the three, or they can elect to reduce the difference, so all randomized fonts are more similar to one another.

While this demo may produce some delightful and surprising combinations, it’s likely that most designers will want to manually fine-tune at least one, if not all, three font styles if they are to be used together. If an individual simply wants to see different stylings in an article-like format, however, this may feel closer to the proper context than casually browsing Google Fonts, which as it turns out, is precisely the library service that FontJoy pulls in for its online tool.

Color Palettes

In addition to fonts and type, color of course has enormous impact on any product design and brand. For years, designers have relied on web-based palette tools to select colors. Many of these tools produce a fistful of flat swatches, modeled after the analog experience of going to a paint store and getting some colored squares.

While this approach allows for some flexibility in the output, such as selecting a monochromatic or triad color scheme, it still leaves most of the heavy lifting to the designer. It often takes a considerable amount of time for the designer to define how these shades and hues get applied and prioritized across an array of components or brand elements.

Enter the machine learning-powered tool Huemint. With its placeholder presets for consumer websites, magazine sites, or logo assets, designers only need to choose the number of colors to be used. Users automatically see these values applied in situ on various page elements, along with multiple color scheme options from high contrast to dark mode, pastels, and more.

This significantly reduces the tedium of manually selecting complementary shades, so the designer can focus their creative energies elsewhere.

Huemint offers tremendous value in streamlining the color exploration process. It may even spark some inspiration along the way, as users can see these hue values in a more dynamic setting rather than merely flat swatches.

Logo Design

Every now and then, a company may choose to redesign their logo. Maybe it’s only a refresh. Maybe it’s an overhaul. Or maybe it’s a completely new company with no mark or existing branding. With their AI-powered logo maker application, BrandMark can accelerate the logo design process.

To generate a logo, the designer must provide three things as inputs:

• An entity name and, optionally, a slogan.
• A few keywords that elicit the overall organization’s vibe, and if none immediately come to mind, some sample ones are provided, which is equally useful for anyone simply curious about the tool itself.
• A preferred color palette type, such as vibrant, pastel, and others.

The output is a combination of text and occasionally a simple illustration, over a flat background in a large horizontal carousel. Users can tap through essentially an endless list, but the real magic happens when an isolated logo suggestion is tapped or clicked. When this happens, the user can scroll down to see the proposed logo superimposed over everyday product types like a shopping bag, a beauty product tube, or a vacuum-sealed food packet, as well as digital screens for mobile devices and laptops.

They even go the extra mile and illustrate the logo on printed collateral like business cards, letterhead, envelopes, and more. For this alone, BrandMark holds quite a bit of promise in providing a more solid sense for how a logo direction may translate across two- and three-dimensional products.

Would any of these logo proposals be considered final art? Presently the bulk of the suggestions need massaging—some more than others—but overall, the tool can help narrow down the general brand direction quite quickly.

Accelerating Creativity with AI

Each of these AI tools can expedite the early stages of the visual exploration process for many designers. These activities may range from working on a brand refresh or building a new design system from the ground up to experimenting with color palettes, font family combinations, and brand style direction.

While much of the output at this time may not be polished enough to be published without some tweaking, these tools have the potential to save a considerable amount of time, especially when compounded across a multitude of projects. As part of a designer’s regular arsenal, these tools may also spark alternate ideas and directions to re-input into the tool as a new prompt or to further refine on their own.

For decades, Orion Innovation has been helping businesses operationalize design across their entire digital ecosystem. Explore our Experience Design capabilities.

Amanda Chan is a Senior UX/UI Designer in New York, driven by a passion for devising scalable solutions with a strong visual impact for clients across a range of industries.