Product descriptions – how to use ChatGPT in an online store?

It won’t write itself

Let’s start with the fact that the use of artificial intelligence-based tools can significantly shorten the time needed to prepare product descriptions. It can also help you save money that would normally be spent on hiring copywriters to write descriptions (as well as finding the right contractors). However, do not expect that all you need to do is enter the product name and ChatGPT will spit out a description, which you just need to copy to the product card. Contrary to appearances, to obtain satisfactory content from an AI content generator, you must first provide it with an appropriate “base”.

Basics of content generation

This paragraph will not replace a course on using AI in e-commerce, but I will try to explain five fundamental terms:

Artificial Intelligence (AI) – this term, used in all cases, refers to the ability of computers to learn and imitate human behavior.

We wrote more about AI in one of our previous articles.

Advanced language models (LLM) – algorithms for processing data and generating content based on it. Today, language models are often associated with AI, when in fact they are only part of the set that includes deep learning and machine learning.
Prompt (sometimes also input) – a command in which the ordering person describes what effect he wants to achieve. Prompts function both in the generation of text and visual content (graphics, and recently also films and animations).
Chatbot – a tool used for communication between humans and computers. Chatbots work on the principle of conversation – a human enters subsequent prompts (queries, data sets), and the computer responds by generating the desired content.
ChatGPT – a tool based on an advanced language model. It processes huge amounts of text and data available on the Internet and then compiles them into content based on prompts.

Advantages and disadvantages of using AI-generated content in product descriptions

Once you know the basics, you can focus on generating content. Artificial intelligence tools do not (yet) replace humans in everything, but they bring many benefits that cannot be ignored. Here they are:

Saves time

Writing guidelines for a content generator takes much less time than writing an entire product description – especially if you expect longer descriptions. For ChatGPT, or any other tool, 2-3 sentences of the prompt are often enough to start working. After a few “revision sessions” you will have the text ready to use.

Less expenses

Content as simple as product descriptions does not necessarily require hiring copywriters for this specific task. ChatGPT still needs to be supported and properly prepared prompts must be provided, but with proper organization it will still be a cheaper solution than creating content from scratch. Even if you need someone to do it for you anyway, they will probably be able to prepare much more descriptions using ChataGPT in a similar amount of time than if they had to write them themselves.

Support for writers

ChatGPT does not have to be a way to reduce employment – you can use this tool to support copywriters in creating content. A chatbot can help overcome writer’s block and relieve content workers from boring and repetitive tasks – so that they can focus on creative, more important tasks for the company.

Multi-language support

Many AI tools handle Polish well, and the vast majority are based on English content. So if you want to prepare content in different languages, without necessarily knowing them well, AI solutions can help you significantly.

Consistent content

ChatGPT is a tool that tries to follow the guidelines. If your prompt contains style tips that you replicate in subsequent prompts, you can expect that the content prepared by the algorithm will follow a similar style. For example, you could include a guideline requiring ChatGPT to include a phrase or term specific to your brand in every product description.

However, there are several aspects of generation that are worth paying special attention to before you commission a chatbot to prepare a thousand descriptions.

Limited creativity

ChatGPT and other content generators are able to create correct texts, but their creativity is limited. Although product descriptions rarely require flair, it’s worth considering – especially if you want to stand out in the way you communicate.

Limited understanding of context

Often, even unconsciously, we refer in our content to current events, the political situation or even famous people. These types of references can pose a serious problem for an AI tool. So if you plan to make your product descriptions a small work of art… you may need much more than a functional chatbot.

Data dependence

Both ChatGPT and other artificial intelligence tools operate based on available data. However, the information that chatbots gain access to does not always have to be reliable. Moreover, there are already known cases of blocking databases from AI robots – precisely to limit their use in commercial content. It is worth keeping this in mind and checking each AI-generated content carefully.

AI not for free

Contrary to the first impression, using content generators is not free. First trials and tests are usually available free of charge, but if you want to use AI in the everyday activities of your company, be prepared to spend some money. It should also be expected that as technology develops and the range of generator capabilities increases, fees will regularly increase.

The status of AI at Google is unclear

Creating content in the context of achieving good results in the Google search engine is a matter that requires a lot of flexibility. Google’s algorithms are not always clear and they change frequently. In 2023, news broke that Google would accept content created by AI (which may have been related to the development of its own AI), but numerous reports throughout the year showed that Google’s priority was still to promote high-quality content, not content in in general.

So if you want to create content that will be presented high in Google results, you need to be up to date with current developments. The latest algorithm update (March 2024) aims to remove low-quality AI-generated content from Google searches. It is estimated that as much as 40% of this type of content will disappear from the results proposed by the most popular search engine in the world.

How to properly create product descriptions in ChatGPT?

Time for specifics. Creating a product description in the OpenAI tool (as well as other text generators) requires entering a prompt. You can do it easily:

Example::

Create a product description for the online store website. The product described is an X brand blender.

This approach will mean that, yes, you will receive a product description, but it will certainly not meet your expectations. Instead of creating laconic instructions, try to include as much detail as possible, both in terms of form and content. Here’s an example.

Example:

Create a product description for the website of an online home appliances store. The described product is an X brand blender, black. The blender is equipped with a glass jug with a capacity of 1.5 l and is distinguished by a five-speed speed regulation. Here are the product specifications.

Power: 1000W

Black colour

Additional functions: smoothie, grinding, pulse mode

The description should be 1200-1500 characters long, including spaces, and written in a professional, factual style. Remember to divide the description into several sections – provide each of them with an appropriate heading that describes it.

The above description, although it could certainly be more detailed, already contains enough data for the generator to be able to create something that will largely meet expectations. An important feature of ChataGPT is the ability to further improve the generated description. If the result obtained does not suit us, we can indicate specific places and elements that need to be changed in order to receive a corrected version of the description after a dozen or so seconds.

Important!

You can “talk” to ChatemGPT, gradually introducing new guidelines into the text. Start with the general idea, and as the conversation develops, introduce further elements (content, form, target group, goal, references to other brands and content, even the use of emoticons or adding quotes) and develop your vision. This way, you can achieve the desired effect in just a few minutes.

The text is ready, is it over?

The final stage of creating content using a chatbot – a stage that cannot be skipped – is manual proofreading and verification of the text. Perhaps this will change in the future, but today content generators still need to be thoroughly checked – especially if the product you want to describe differs from the others in technical nuances.

What should you do at the end of working with the text?

Check the text with anti-plagiarism software – although Google’s approach to the use of AI-generated content is not entirely clear, there is no doubt about plagiarism – content containing copied fragments of already published texts is “buried” and does not reach high positions in the search engine. So make sure the chat hasn’t accidentally copied parts of the text from somewhere else.
Check if everything is correct – the chat may “confuse” some product features, add something of your own or include information about a different model in the text.
Check the text length – although it may seem a bit strange, chat is not always good at… counting. So if you set an expected number of characters, don’t assume that this command will be 100% fulfilled. Check in your text editor whether the tool has coped with this difficult task.
Carry out linguistic proofreading – as after every chat, the text should be carefully checked for linguistic correctness. Pay special attention to punctuation and inflectional endings – you are unlikely to see any spelling errors made by the AI.

Only ChatGPT?

In the Polish (and probably not only Polish) discourse, the OpenAI product, i.e. ChatGPT (currently in the GPT-4 version), has become somewhat synonymous with content generating programs that operate using generative language models, a bit like Adidas refers to a certain type of sports footwear. In reality, however, there is quite a lot of competition in this field. Tools such as Copy.AI, Frase, Jasper or INK are becoming more and more recognizable and may constitute – if not equal competition – an interesting alternative to the most recognizable giant.

When looking for the best solution for creating product descriptions, it is worth considering these alternatives – especially since some of them offer solutions supporting SEO – for example, tools such as Product Description Generator from INK.