Site icon Ringostat blog: posts about marketing and analytics

Test drive: how and where to generate creatives for advertising using AI

The cover of the American magazine Cosmopolitan became one of the first bold and striking examples of DALL-E neural network application in design. Since then, the world’s companies have been actively experimenting and surprising users with exciting product design, advertising, and communication approaches. But is it suitable for business? Iryna Harmash, the content marketer at Ringostat, investigated different AI services for image generation. Find out the advantages, disadvantages, and peculiarities of creating creatives for advertising or social media.

Advantages of AI for design creation

We’ve all seen incredible artworks created by artificial intelligence — the high level of execution, bordering on real art, is impressive. At the same time, we have witnessed scandals related to generative AI more than once. For example, the funny story of the Pope’s white down jacket. Sometimes, there are also loud calls to ban or at least suspend mass access to technologies, given the underdeveloped legal framework.

Are such controversial cases a reason to reject the benefits and opportunities of AI for business? In our opinion, no.

At Ringostat, we use AI capabilities in the work of our marketing team. First, we use it to inspire and generate ideas, as well as to increase productivity in working on textual content. So now we decided to investigate what tools and platforms can be useful for creating creatives.

Another equally important advantage of AI is the ability to quickly and easily create images for your own needs. If a company doesn’t have its own artist or photographer preparing the necessary graphic materials for advertising, marketers have to use photos and illustrations available online. Often, these are images from stocks or photo banks. And there’s no guarantee that another brand, such as a close competitor, won’t use the same images.

For example, to advertise Ringostat virtual telephony, we often use photos of call center operators wearing headphones with a headset. But, firstly, there are a limited number of such images, and secondly, our competitors are also hunting for similar photos. And even if the same faces of the models are played up in different ways in branded colors and styles, confusion is not excluded. And the recognition among potential customers is reduced.

Artificial intelligence helps to avoid this problem. AI users can freely experiment with the shape, style, color, and other elements to create a unique ad.

Disadvantages of AI-based image generators

The first drawback faced by almost everyone who tests the capabilities of artificial intelligence to create images is imperfect images of people

For example, here’s a great result with a rather photographic face of girls and, unfortunately, unnaturally long arms 😀

But this won’t necessarily be a hindrance if the specifics of your business don’t require demonstrating how a person uses the product. For example, if you sell furniture, toys, flowers, etc., you can demonstrate such products directly in your ads. An attractive image of an appetizing dish may well work even better than a smiling human face. Below are some samples of advertising banners created by AI without human images.

An image of sushi generated by DreamStudio. Below in the article, you will see a banner created on the basis of these illustrations

Another disadvantage of using AI is the limited functionality of common services. Although creating a unique image from scratch is a fascinating activity, it is not enough for working on advertising materials. You need more functionality: adding text, logos, buttons, and graphic elements. But not all platforms provide such features.

First, we’ll look at the platforms that include a range of features: 

image creation and tools for further work on the banner. And then, as a bonus, we’ll look at a few platforms that offer only one thing: generating images or helping you work on promotional materials. All the services listed in this article have a common feature: they are powered by artificial intelligence.

But first, let’s talk about how to properly formulate a prompt, i.e., a request to artificial intelligence to get the desired result.

🔥 Get more leads and boost your sales

Ringostat will show you how to optimize campaigns and how your sales reps work

How to write a prompt to generate an image

  1. Be clear and specific. For example, instead of the text query “monster”, try sending a query “fluffy cute monster with sunglasses”. 
  2. Specify specific image parameters: size, color palette, and type.
  3. Add keywords to make your request more specific. For example, if you want to create a banner about traveling, the keywords can be a bus, palm trees, outdoor pool, sunglasses, cocktail, and maps.
  4. Use simple sentences. AIs understand laconic phrases and sentences without complex grammatical structures better. For example, instead of writing “A picture of a city with many tall buildings glowing in the night sky,” it is better to write “Night city, tall buildings, neon glow.”
  5. Use examples if the service allows you to upload your own source image. This will help AI better understand what you want.
  6. Refine your query using built-in settings and filters. For example, some services allow you to choose an artificial intelligence algorithm – Stable Diffusion, Dall-E 2, etc. Or you can specify the desired image size right away.

According to the recommendations of the NightCafe Studio service, the text of the prompt should be in sequence:

Important. The order really matters, so it’s better to mention the central theme or main subject first.

Here’s an example: a deer (main object) passing through an abandoned neon city (auxiliary object), by Dan Mumford and Victor Ngai (names/styles of artists), 8k resolution, a masterpiece, 35mm, hyperrealistic, hyper maximalist (photo/art), crimson, jade (colors).

And here is expected result:

Source: NightCafe Studio email newsletter

Some services offer users manuals and guide with recommendations on how to write a promo. For example, DreamStudio recommends specifying as many details as possible in the request: the subject, style, lighting, tone, etc. The more details you provide, the better the result will meet the user’s expectations.

Some platforms leave a ready-made example in the field for writing a prompt that you can follow. Sometimes, an additional field is added where you can specify a negative prompt — undesirable details or features.

An example of a DreamStudio service prompt

Tip. Like high-quality images that meet your request and dislike results that do not satisfy you. The program will not be “offended”, but on the contrary, thanks to your feedback, it will learn faster and work better in the future.

2 in 1: platforms for image generation and layout of advertising materials

Free AI Image Generator by Canva

Canva is one of the most widely used and well-known online design tools. 

It offers many ready-made templates for creating designs, as well as access to a huge library of graphics, photos, and fonts that can be combined and customized to bring your vision to life.

Recently, Canva’s functionality has been expanded with artificial intelligence for image generation. Now, users can not only choose from ready-made graphics but also create artwork based on a text sample. And immediately use the generated image in their design.

The feature “Text to image” has appeared at the bottom of the toolbar on the left of the page. Click on the purple icon, and you’ll see a field for entering a prompt and several options for the style of the future image. You can also choose the desired format — square or rectangular in portrait/album orientation.

Keep in mind that you can choose the format only for the generated image, not for the banner as a whole. In the free version of Canva, you can only create square banners.

In our experience, human figures in generated images have obvious flaws. If your query is complex, it is likely that you will not get the desired result the first time. However, AI processes short and simple samples better.

If one of the images meets your requirements, you can immediately create an advertising banner with it. All you need to do is use Canva’s free functionality:

To create an advertising banner, you can use one of thousands of free templates as a basis. Or you can create one from scratch, as we did, using a generated image and available functionality:

Price 

Canva’s image generation tool is available for free to all users. There are no limits. The platform also provides access to a large set of free tools: templates, photos, graphics, etc.

Advanced features, such as customizing the format of an advertising banner, become available after subscribing to one of two paid plans, which cost from $100 per year.

🤝 Build effective communication with Ringostat telephony

Connect a virtual PBX in one working day and receive calls anywhere in the world

Microsoft Designer

It’s a great platform that has full functionality for creating designs: ads, invitations, logos, social media posts, website banners, and much more. You can start with your idea or choose from thousands of ready-made templates and customize them to your liking. You can also create your own image using AI.

Microsoft Designer uses various AI algorithms:

Microsoft Designer is in beta testing, at the time of writing. The program works in web browsing mode, but to get access to the functionality, you need to submit a request and wait for approval for about a week.

The tool shows examples of successful prompts. Move your mouse over one of them on the right side of the page to see a prompt of that image. You can click on the image you like to create something similar. Or you can enter your own prompt in the empty box on the left:

It is still difficult to depict people naturally: body proportions, faces, and fingers. And especially in combination with such complex details as, for example, headphones and headsets. However, if you experiment with the software, you can get satisfactory results.

The prompt: sales rep, wearing headphones with a headset on his head, at a computer, against a light background

We also decided to create ads for imaginary products, such as cosmetics, for a change. We got something like this for the prompt “a facial care cosmetics based on natural ingredients, photorealistic image, pastel colors”:

After we chose one of the generated images, we were immediately offered several ready-made templates for creating a design. Moreover, we could choose the format we wanted:

Interestingly, the proposed designs already contain the key phrase from your prompt instead of “Lorem Ipsum”.

We chose one of the options and tried to create our own banner based on it. Moreover, you can change the template while working. The “assistant” will constantly add new ideas for inspiration (right):

Opportunities:

A banner to advertise a fictional brand and cosmetic product, which we created in 45 minutes from scratch:

I like the fact that the tool offers completely different templates each time, based on the subject matter, color, and style of the original image. For example, for the query “sales rep, call center operator, wearing headphones with a headset on his head”, the following banner templates are offered:

But AI offered completely different ideas for a poster or story design on the topic of yoga:

Not to mention the smart formatting, alignment, color matching, and other “magic” of working with templates.

Price

The developers promise that Microsoft Designer will be free after the release in a web application format, albeit with certain functional limitations. Additional features will be available for Microsoft 365 subscribers, which costs $69.99 per year.

Adobe Express

It’s a powerful tool that can be used to create banners for social media, marketing campaigns, presentations, or anything else you might need.

Firefly is an image generation tool from Adobe that is currently in beta testing. To test its capabilities, you need to submit a request and wait for approval. Unfortunately, at the time of writing, we have not yet received an invitation, although we submitted a request two weeks ago. So we can’t tell you how the tool works and show you examples.

Instead, the Adobe Express banner creation platform with a variety of different templates is available to everyone:

The workspace provides all the possibilities for working on promotional materials:

So, we choose one of the simple free templates. Based on it, we create a post for Facebook using the creative that we previously generated with the help of artificial intelligence. Just like you can do after receiving creative from Firefly.

We changed the text, colors, and the main image. We added round elements and icons. Here’s what we came up with:

It is worth saying that the web program is slow. It took us over an hour to create the very simple banner you see above.

There are also drawbacks related to the interface. For example, when you select any element, additional editing fields are constantly opening on the right and left. They cover the workspace and get in the way, especially if your screen is small.

If you’re used to Figma, for example, you may find the lack of some options inconvenient. For example, selecting two elements at the same time or undoing a previous action with Ctrl+Z. But this is a matter of habit. 

Moreover, in contrast to the disadvantages listed above, the platform has advantages:

Price

Adobe Creative Cloud is free by default. It gives you access to thousands of templates, resources, and fonts. You get basic photo editing and effects to help you create the images you want. You also have 2 GB of cloud storage at your disposal.

There’s no clear information yet on whether Firefly will be free after the beta. Here’s what the Adobe developers have to say: “Once Firefly is out of beta, creators will have the option to use content generated in Firefly commercially.” This may mean that Firefly will be available only to Creative Cloud users or for a fee. 

The premium plan costs $10 per month and includes thousands of additional templates, as well as other AI banner features. In addition, you get access to more than 160 million free Adobe Stock photos.

⚡ Provide yourself with high-quality and reliable communication with Ringostat telephony

Make and receive calls with a couple of clicks — all you need is a laptop with Internet access

 

Use dozens of options that save managers time and wow customers

 

Rely on technical support that responds in 4 minutes

Pixlr X

It is a free online service for editing photos, creating static designs, and animations using AI. 

The multifunctional platform offers the following advanced features:

First, we tried to generate our own image to evaluate the tool’s capabilities.

To make the creation of a promo image as simple and convenient as possible for users, the platform offers tips on type, style, and composition. It is enough to enter a short description of the desired object in the blank line and click the buttons with additional image attributes to complete the promo.

The resulting image can be further refined by clicking the “Variations” button:

If you are satisfied with the final version, you can go to the “Editor” section and start creating a banner. The following options are available here:

Below is an example of a banner that we managed to quickly create using an AI-generated image of a chameleon and the functions, fonts, and graphic elements available in the free version.

So, the basic functionality and free graphics are enough to create simple banners. The service updates the limits daily and provides five credits for generating images.

At first, the program’s interface and the arrangement of tools may seem a bit unusual. For example, we didn’t immediately notice how to add common shapes to a banner: circles, rectangles, etc. But we eventually found all the necessary settings.

Important. When working on a banner, it may seem that all graphic elements, including fonts, are very “pixelated”. But if you choose high quality while saving your banner, everything will be fine in the end.

Price

Most of the functions for working with banners are available for free. However, the ability to generate images in the free version is limited. The user is given only 5 credits at a time. The limits are updated weekly.

⚡️ Increase the number of leads with callback Ringostat

Get 30% more calls from the site, without additional investments in advertising and SEO

Visme

Visme is an online platform for creating presentations, infographics, designs, and videos. It provides access to thousands of ready-made templates and allows you to create your own designs. Visme also has an AI Image Generator feature for generating images based on text queries, which is still in beta testing.

To create a design, choose one of the proposed templates. For convenience, they are grouped by purpose: presentations, and posts to various social networks, including carousels, infographics, banners, etc.

To edit the selected template, there is a toolbar on the left. If you click the “Photo” tab, a field opens with a line to search for a photo in the service database, upload your own image, or generate a new one using AI. The latter option is available under the blue Generate AI Image button at the bottom of the panel.

You can choose from styles:

Frankly speaking, human faces are imperfect here. But the tool did a better job of creating images of objects such as pizza, wedding bouquets, and jewelry:

Visme has special functionality that other popular services do not have. For example, you can:

Price

Visme has a free plan that includes limited access to templates and materials. You will also have 10 credits for generating AI images. If you want more features and functions, you can choose from two paid plans: Personal ($15 per month) and Business ($29 per month).

Tools for generating images using AI

Now let’s look at tools that can only create images on demand. To further process them and turn them into advertising banners, you need to look for another service.

DreamStudio

The main feature of the platform is the high realism of people’s images. This is important if you need to create such advertising banners. 

Another advantage of DreamStudio over other platforms is that it generates not only square creatives. You can choose horizontal and vertical formats in different proportions. 

You can also immediately request the number of creatives you need — from one to ten. Unlike other services that generate two or four images by default.

The style is also diverse. In addition to the photorealistic image style, there are other quite interesting ones, for example:

In the prompt, we asked for a creative with “call center operators smiling but focused on their work, wearing headphones with a headset.” We also specified the desired style – “photographic”. We also took advantage of the opportunity to request more images and chose a different image format than AI usually offers — rectangular. And we got the following result:

AI generated the operators we needed quite realistically:

Unfortunately, the platform doesn’t have any tools for further work with the generated image, and even editing it is not available in the free trial version. So we had to upload the file and work with it separately. After choosing one of the options, we sent it to Figma, where we already had our own ready-made templates with text and logo.

All we had to do was slightly adjust the colors of the banner elements to match the image created by the AI. And here’s the result:

Advantages:

But you can see aesthetic defects in human appearance in some pictures. The saturation and contrast of tones will also seem exaggerated to some.

Tip. If your ad is targeted at people from certain countries, you should specify the appearance features you need in the promo: European, Asian, etc. This way, your ads will be better perceived by your target audience.

Price

The payment options are quite flexible: you can pay for the required number of images. For example, 100 credits are available for $1, and 1000 for $10.

Image Creator by Bing

Bing is a search engine owned by Microsoft. Image Creator by Bing is an additional functionality for generating images based on text queries in the Bing chatbot, available for free to all users with a Microsoft account. 

The system is powered by OpenAI’s DALL-E neural network and uses a library of online images and texts for training.

To use Bing Image Creator, you need to go to the site and sign in with your Microsoft account. Image Creator is also available as a built-in app in the Edge browser:

New users are given the opportunity to create 25 images with acceleration, i.e., in 10-15 seconds. The generation of subsequent images may be slower, but new “speedups” are provided every week. 

It’s easy to use: enter any text description of the image you need and wait for the AI to generate four square images according to your request. 

Unfortunately, we were not very satisfied with the images of smiling call center operators. Their faces are too “artificial”.

But for a simple prompt such as “a festive gift wrapped with a bow, dark background”, the AI did a great job:

We used one of the four images to create an ad using our own template in Figma in 5 minutes:

The developers warn about the possible shortcomings of future images. And, unfortunately, the images of people are not suitable for use in advertising, according to our experience.

Please note that each image is marked with a watermark in the lower left corner.

Price

The main advantage of the tool is that there are no fees and no restrictions on the number of requests. According to the developers, the aim of the tool was to create an accessible tool for creativity using AI for everyone.

Starryai

Another service that creates unique images using text prompts or uploaded photos. Starryai uses OpenAI’s DALL-E neural network and offers to create designs in one of two AI styles: art or photo.

The service is available as a website and a mobile app for iOS and Android. To use the service, you need to log in to the website or app.

In the free trial version, only one image format is available — a 4:5 vertical portrait. You can also add your own source image, which will be used for AI.

First, we checked how the tool generates an image of a person. The result is quite decent:

However, when the prompt asked for images of a group of people, the quality of the generated images was already disappointing. Faces and bodies were distorted. So, the tool is still good at creating portraits of singles only.

Price

The free trial version includes only 5 credits. Subscription rates range from $8.99 to $63.99 per month, depending on the number of creatives you need, as well as additional features: styles, formats, sizes, etc.

NightCafe Studio

NightCafe Studio is an AI-powered image creation and sharing platform. 

NightCafe Studio uses the following algorithms to generate images:

Each of them has its own features and capabilities. For example, the former creates realistic images with high detail, while VQGAN+CLIP creates abstract and artistic images. Thus, the tool allows you to create images in a variety of styles, including hyper-realistic, anime, 3D, imitation of painting, etc.

We get such result by prompt “a single pastry with almonds, honey, and chocolate, a light background, a photorealistic image, pastel colors”:

Unfortunately, the realistic portrayal of people here is not so good. The prompt sounded like this: “a man, a call center operator, working at a computer with headphones and a headset, photorealistic”. Here is the result:

NightCafe Studio works as a website and a mobile app for iOS and Android. The service also has a server on Discord, so if you’re more comfortable using the tool in a chat room, join and create there.

Price

It is conditionally free. Each user receives 3 credits when registering and 3 more when subscribing to the service’s newsletter. Five new credits are added daily, and there are other additional conditions for increasing the limits. 

Paid subscriptions are available on a variety of terms, depending on the required amount of content generation — from $4.79 to $39.99 per month.

Airbrush

The service offers:

For example, for the query “call center operator”, we see such images with text descriptions for each of them:

This service uses the advanced GPT-3 and CLIP algorithms to convert text to images and vice versa, as well as improve the quality of the result.

The following advanced settings are available for image generation:

Interestingly, in addition to generating images on demand, you can also get the opposite result here — turn an existing image into text. This can be useful for creating and further using successful promos.

The platform also provides tips on how to work with the functionality: how to improve the quality of images, how to write a promo, etc.

Price

Airbrush offers two tariff plans:

You can also get an additional 20 free credits by leaving a review of the service on Google Play, Capterra, G2, etc.

Stable Diffusion Online

This is a free image generator that works without restrictions. You don’t even need to register. Just enter the prompt in the appropriate field and click the “Generate images” button. The service also doesn’t store users’ text or images.

The Stable Diffusion algorithm is used to create images.

Images of people are not very good. But objects and abstract patterns are better. For example, here’s a picture that the service created for us based on the query “fresh buns, drizzled with honey, on a light background, view of the sky, pastel colors”:

But the service also has drawbacks that cannot be ignored:

However, the service promises to improve and expand the functionality over time, which it is actually working on.

Constructors for creating banners using AI

Finally, let’s look at services that also position themselves as AI-based platforms. They can’t generate images (some of them can’t yet), but they have the necessary functionality to create ads: ready-made templates, tools, and the ability to add your own graphic materials.

Artificial intelligence is “invisible” here. It is used, for example, to:

These handy services can be used to further process the artwork you’ve created before.

Vista create

Vista create is a free and friendly service that helps you create professional designs using ready-made templates or your own creativity. 

Here you can:

Here are, for example, suggested templates for Facebook ads, each of which can be modified to suit your own needs:

The service also offers detailed tutorials explaining each stage of the process.

Price

Most of the tools in the Vista create platform are available for free. Premium, available for 14 days for free, costs $10 per month and provides additional features. For example, teamwork, unlocking additional templates, and advanced functionality.

Abyssale

This service also does not have an image generation function. But the platform uses AI to simplify and speed up the design process.

We liked that no design skills are required to create advertising banners. There are whole sets of templates already made in the same style, but in different formats — for different social networks, desktop and mobile website pages, and printed materials. 

You can choose a template from over 100 options. And edit it to suit your own needs:

For example, we used one of Abyssale’s free and advanced templates, which has more than 10 different banners available:

Based on the template, we created the following Facebook ad banner using one of the illustrations previously generated by DreamStudio’s artificial intelligence:

Price

The rates for regular use of the Abyssale service range from $49 to $299 per month.

Snappa

Another quick and easy online tool for creating various graphics for social media, advertising, and blogs.

There’s no image generation tool, but Snappa does use AI for some functions, such as background removal, filters and effects for images, etc. 

These features allow you to easily create professional designs without complicated settings and edits. You can choose from thousands of ready-made templates or create your own design from scratch using the available features.

Here is an example of a simple banner created in 20 minutes based on one of the free templates with minor changes:

Price

Unfortunately, the free version of the service has limited features. Although each user is offered 6,000+ templates and 5,000,000 photos and graphics, as well as full functionality for creating designs, only 3 ready-made banners per month can be downloaded. Unlimited downloads are available only in paid versions of the service, which cost $10-15 per month.

Conclusions

  1. When working on creatives, you shouldn’t rely too much on the “magic” of artificial intelligence. To create truly informative and attractive banners, you still have to work hard and, as they say, “play with fonts.”
  2. To get the desired result right away, you need to write a “correct” prompt, i.e., a query. It’s important to specify the desired item right away, then add essential features and characteristics, and specify the style and colors, if necessary. Some services offer their own tutorials on how to write a successful promo.
  3. Some platforms, such as Canva, Microsoft Designer, Adobe Express, and Visme, have powerful functionality. In addition to generating images, they offer a full range of design options, including graphics, fonts, color schemes, logo uploads, and much more.
  4. There are also services that can only generate images. You can use them for inspiration or to find references or continue working on the created image in other programs.
  5. There are also design programs based on artificial intelligence. These are, for example, Vista Create. Snappa, Abyssale, and many others. They can’t generate images, but AI invisibly helps to quickly and easily create attractive designs based on ready-made templates and convenient tools.
Exit mobile version