Artificial Intelligence (AI) technology has been revolutionizing nearly every industry in the tech world, including website design. AI has the power to do far more than humans can do, whether it’s crunching massive amounts of data, coming up with a handful of ideas in mere seconds, or generating missing pieces of an image.
Although AI has the potential to improve and enhance various aspects of web design, it does need to be approached with caution. There is a right and a wrong way to implement AI in your web development process, especially where retail websites are concerned.
If you’re a designer who creates websites for retail businesses, and you want to start using AI in your workflow, make sure to keep the following best practices in mind.
Now that tools like ChatGPT and Google Gemini are mainstream, it seems like everyone is using them to generate content for their website, like product descriptions, regular blog articles, sales copy, and even meta tags. If you’re like most designers, your clients probably ask you to create at least some kind of content for them. Using AI to create copy will save you time, but if you publish it as-is without making sure it’s accurate, you might end up publishing incorrect information that will make your client look bad.
Your web design clients depend on you to execute their project flawlessly, and while they would probably be forgiving of simple mistakes, don’t risk making mistakes that can be prevented.
Imagine using AI to create a blog post for a retail client who sells Skateboards, and you end up with content that names several fictitious famous skateboarders who don’t exist. If you’re not passionate about skateboarding, you might not know the names are fake. However, your client will notice, and they’ll wonder where you got the names. If you don’t want your client to know you’re using AI to create their content, this type of mistake will let the cat out of the bag.
Even if your client doesn’t mind you using AI, it’s crucial to verify every piece of data in your generated content. Published mistakes will be seen as a reflection on your client and can be off-putting to their customers.
Many business owners have found out the hard way that it’s too expensive to pay for photo shoots, whether it’s just a product shoot or one that involves human models. Professional photographers are expensive no matter what they’re photographing, and that has made business owners turn to AI for cheaper, easier solutions. As a web designer, you might be asked to generate some AI product mockups for your clients who want to save money.
Generating mockups is fairly easy. You can use generative AI tools, like Midjourney or Ideogram, to help your clients bypass the hassle of a photo shoot. If they’re selling t-shirts, for instance, they don’t need to get models to wear their shirts and then hire a photographer. With the right prompts, you can create realistic images of people wearing blank t-shirts in the setting of your choice. With some Photoshop skills, it’s not hard to change the color of those t-shirts and place your client’s designs on top to create realistic mockups. However, AI-generated mockups can be misleading.
While AI-generated mockups seem like the ultimate solution to expensive photo shoots, it can work against your client’s intentions. Some customers won’t mind if the shirt in a mockup photo isn’t the exact brand they receive, but many will notice and be upset over the differences. There are a lot of people who make purchase decisions based on visual cues, like sleeve length, collar width, whether or not there are side seams, and other elements that can be visually discerned from product photos. If customers think an AI-generated image depicts a model wearing a real t-shirt, they might buy it based on characteristics that the actual product doesn’t possess.
If your web client asks you to create AI-generated mockups for them, make sure they understand how important it is for those mockups to accurately depict their product. If you are on the hook for any type of results for a client, you may want to discuss the matter in-depth with them to help them understand the potential issues before they make a decision.
Compared to stock photos, which can cost $3-$30 each, AI is much cheaper. You can pay around $30 per month for a subscription to an AI tool and generate an unlimited number of images. Financially, it makes sense to generate images whenever possible, especially if you include those images in the cost of your development fees.
AI-generated images can be an amazing replacement for boring stock photos, but you need to be careful when using them in place of stock photos. For instance, AI is notorious for producing hands with thumbs on the wrong side and the wrong number of fingers.
Also, sometimes limbs are duplicated, twisted, or in the wrong proportion. Some of these mistakes are easy to fix in a graphic design program, but many are not.
Always double and triple check the images you generate for your clients to ensure they look realistic and accurate. It’s not just people that can appear messed up, so pay attention to every detail. For instance, if you generate an image of a fire burning in a fireplace, make sure all the logs are actually located inside the fireplace. You’d be surprised what can slip by if you don’t look carefully.
You might be surprised to learn that there are laws that govern photo retouching, and that can cover both AI-generated images and images you retouch with AI. You probably already know that in the United States, the FTC requires businesses to accurately represent their products, but there’s more to it than that.
If you work with international clients, you might be required to adhere to much stricter laws regarding images depicting models. For example, in France, photos that have been altered to make a model’s silhouette narrower or wider must be labeled “photographie retouchée.” There are similar laws in Israel, and other countries.
If you take existing images and use AI to retouch them, your client needs to know so they can label the properly. The consequences for your oversight in this situation will fall on your client.
If you don’t work with international clients, stay on top of proposed legislation in the U.S. because these kinds of laws could be enacted in the U.S. in the near future. It’s also possible that the existing advertising laws could apply to AI-edited images if they are altered in a way that misrepresents a product or depicts a model with an unrealistic body. You never know when there might be a Supreme Court ruling that changes everything.
Since AI-generated images aren’t eligible for copyright protection (as of now), anyone can use the images you generate using programs like Midjourney, Kittl, Playground, and Ideogram. By default, these applications make your generations public, but you can change your settings if you subscribe to a high enough paid plan.
It will almost always cost more to ensure your AI-generated images remain private, but you won’t have to worry about anyone else finding your images and using them for commercial purposes. It could become an issue if you generate specialty images for a client and one of their competitors finds your images and uses them in their ads or on their website.
If you’re not much of a writer, but your web design client is asking you to just use AI to generate a bunch of copy for their blog, try to talk them out of it if you can. There are plenty of ways to use AI-generated copy, but it’s not currently good enough to rely on for blog articles that will keep a visitor’s attention.
Business owners that have blogs for search engine optimization (SEO) purposes sometimes forget the importance of publishing content specifically for their audience. This requires a custom, human touch that goes far beyond the general nature of AI-generated content.
If you’re not keen on writing content for your web clients, be prepared to refer them to a content writer you know, or politely decline the request. If your client understands that AI-generated content isn’t likely to engage their market enough to persuade them to join their email list or buy something, you can do it if you feel like it’s worth your time. However, if you’re a designer who likes to ensure your clients get the best results possible, recommend hiring a real writer first.
These days, many web designers also become content writers since clients often prefer working with one person who can do as many tasks as possible. If a retail client asks you to write sales copy, that’s not a job for AI. Writing sales copy isn’t the same as typical blog content in the sense that it needs to be specifically crafted for increasing conversions. That isn’t something AI can do very well and it’s actually a task best suited for a professional copywriter.
AI tools are amazing at creating general copy, including full blog posts on basic topics. However, don’t rely on it for copy that is supposed to influence visitors into taking specific actions. AI-generated sales copy might see some success, but it will never get high conversion rates on par with intentionally-crafted persuasive copy.
If you are tasked with writing sales copy for a web client, make sure you aren’t going to be held responsible for conversions before you entertain using AI to write those pages.
At least once in your life, you’ve probably searched the web for statistics and found a bunch of sites quoting the same numbers, but the original source was nowhere to be found. Sometimes statistics aren’t even real and people just repost the information assuming it’s legitimate.
Generative AI systems have been trained by the entirety of the internet, which is why they often make so many mistakes. Never rely on tools like ChatGPT to provide you with statistics or even market research data. You have no way of knowing the sources that are being used to provide you with results and these tools won’t give you any links to original sources.
Whether you’re using AI to retouch photos, replace stock photos, generate mockups, or generate simple content, use it wisely. It’s a useful, yet imperfect technology that still requires human oversight to ensure accuracy. Your web clients might not understand the limitations of AI, so when you’re tasked with using this technology in a project, make sure to use it responsibly and ethically so the final result reflects positively on your clients.
.svg)
