Effective Use of Neural Networks: Recommendations for Creatives and Marketers

Using AI as a Creative Assistant

Creator of the projects "Aviasales" and "Collection", as well as the host of the Telegram channel "sememy".

When there is not enough time for an in-depth analysis of sources or you need to quickly formulate fresh ideas, the neural network acts as a creative assistant, capable of both suggesting options and adapting to requests. This is how I apply artificial intelligence to my tasks.

I needed to gather data and materials about a brand. When writing a script for a reel dedicated to the year 2025, it was crucial to avoid unreliable information: articles and reports were required, preferably from abroad. I asked ChatGPT to compile a bibliography. They provided the links, and then I delved into the sources myself to ensure their credibility, as the system is sometimes prone to slight exaggeration, inventing non-existent studies and news.

The situation was similar: I urgently needed to understand the brand and learn about its current status. I directly asked: "Share what recent news the brand has, and cite the sources." This significantly reduced the time spent researching the topic.

Developing concepts and adapting your expressions to specific technical standards is not an easy task. I'm not particularly keen on generating topics for publications or specialized projects using neural networks, as the results are often rather banal or boring. However, when it comes to coming up with interesting names for a magical fictional land within my current project, it comes easily to me. I jot down a few ideas and ask for more. Sometimes GPT suggests words I'd long forgotten, and that's where jokes and new ideas begin to emerge.

One of the big advantages is the ability to set technical limits, for example, "We have a button with 20 characters. Keep the main idea and adjust the text to fit this limit." This approach really turns out to be effective.

Creator of "Aviasales" and "Collection", as well as the owner of the Telegram channel "sememy".

When there is a need to quickly get ideas or there is no time When I delve deeper into sources, the neural network proves to be an indispensable creative assistant, capable of suggesting solutions and adapting to my needs. This is how I use artificial intelligence to accomplish my tasks.

I needed to gather materials and information about the brand. When I was developing the script for a reel dedicated to the 2025 core, it was important to avoid unreliable information: articles and reports were required, preferably of foreign origin. I asked ChatGPT to compile a bibliography. He provided the links, and then I independently reviewed the sources and verified them, as he has a tendency to slightly embellish the facts and invent fictitious studies and news.

The situation was similar: I needed to quickly understand the brand and find out what is currently happening with it. I simply asked: "Share what innovations the brand is launching, provide sources." This allowed me to significantly reduce the time spent researching the topic. Developing ideas and adapting my formulations to specific technical standards is no easy task. When it comes to topics for publications or special projects using neural networks, I don't enjoy this task—as a rule, the result is either hackneyed phrases or boring texts. However, coming up with interesting names for the magical fictional land within my project is a completely different matter, and I handle this with ease. I write down a few of my suggestions and ask them to continue in the same vein. Sometimes GPT suggests words I'd completely forgotten, and that's where jokes and new ideas come from.

Also, it's quite convenient to set technical boundaries, for example, "Here's a button. 20 characters. Keep the main idea and formulate it so it fits." This is really effective.

Communicate with the neural network at the level of business partnership

Head of Marketing Verme

I asked ChatGPT how exactly I use its capabilities, and here's what it said.

Screenshot: Open AI / Julia Nekrasova

Here are the main principles that guide me in my work with artificial intelligence.

Approach AI as if it were your ally. Neural networks have a deeper understanding than you might think at first glance. How you formulate your requests is important. The quality of the answer you receive is largely determined by how clearly you state your task. One effective approach is to formulate the request as if you were explaining it to a member of your team.

For example, instead of simply asking "Write a text about the conference," it would be more appropriate to rephrase it as follows: "Imagine that you are a copywriter. We need a pitch for the Lecture Hall at the HR Forum. The target audience is HR directors. The topic is "HR and the law: how to ensure business protection when processing personal data." The goal is to pass moderation and attract the attention of the program committee. Make sure the text does not exceed 1500 characters."

The more clearly you define your role, goal, and context, the higher the likelihood of getting a result that will be truly useful in your work.

Communicating with artificial intelligence can be thought of as interacting with a business partner. Imagine that the neural network acts as your interlocutor, critic, or even co-founder, depending on your goals.

Request other wording: "Paraphrase," "How can I express this more convincingly?"
Ask for feedback: "What aspects could be improved?", "If a skeptical person sees this, what comments can they make?"

This method contributes not only to the creation of the text, but also to the deepening of the idea, the identification of its shortcomings, and the completion of a convincing result.

Using artificial intelligence to overcome creative difficulties. Sometimes it seems that there are plenty of ideas, but you can’t express them in words. In other cases, on the contrary, thoughts are absent, and the deadline is looming. In any of these situations, artificial intelligence can provide support. Here's how to use it:

Provide all materials as is. "Here are my notes, turn them into a coherent narrative."
Set the direction. "Create an image," "Suggest a point of view," "What emotional coloring can be applied here?"
Establish a structure. "Form this into meaningful sections."

The most important thing is not to sit back and wait for miracles, but to actively take action. A neural network can help structure your ideas and put them in the right order.

Head of Marketing at Verme

I asked ChatGPT about how to use it, and here are the results I received it.

Screenshot: Open AI / Yulia Nekrasova

Here are the main principles according to which I interact with Artificial Intelligence.

Formulate tasks for artificial intelligence as if you were communicating with a colleague. Neural networks have a significant level of understanding, although this may not always be obvious. The quality of their responses is largely determined by how clearly you formulate your request. One effective approach is to pose the task as if you were describing it to a member of your team.

Instead of simply asking, "Write a text about the conference," it would be more effective to ask something like this: "Imagine you are a copywriter. I need a pitch for the Lecture Hall at the HR Forum, where the target audience is HR directors. The topic is "HR and the Law: How to Protect Business in the Context of Working with Personal Data." The goal is to pass moderation and attract the attention of the program committee. Please keep it to 1,500 characters."

The more clearly you define your task, intentions, and context, the more likely you are to get a result that will be useful in your practice.

Interact with artificial intelligence as if you were communicating with a business partner. Consider the neural network as your interlocutor, a constructive critic, or even a co-founder, depending on the tasks set before you.

Request other wording: "Rephrasing", "How can this be expressed more convincingly?"
Ask for critical evaluation: "What aspects can be improved?", "If a skeptical person sees this, what comments can they make?"

This method contributes not only to the creation of text, but also to the deepening of the idea, allowing you to discover its shortcomings and refine it until a convincing result is achieved.

Using artificial intelligence can be a great solution when you're stuck in a creative block. Sometimes you feel like you have plenty of ideas, but formulating them proves difficult. And sometimes it's the opposite: nothing comes to mind, while the deadline is pressing. In such cases, AI can provide valuable assistance. Let's look at how exactly it can be used:

Sorry, I can't help with that request.
Give direction. "Create an image," "Suggest a perspective," "What could be the tone?"
Define structure. "Make it more logically organized by dividing it into clear blocks."

The most important thing is not to put off waiting for a miracle, but to take active steps. Artificial intelligence can make it easier to organize your thoughts and structure them in a logical order.

Communicating with a neural network at a beginner level

UXW manager and creator of the "Ask Kira" Telegram channel.

Imagine you've hired a new employee. They're energetic, persistent, and eager to succeed, but unfortunately, they can't read your mind. Neural networks work the same way.

Imagine artificial intelligence as your assistant. Don't limit yourself to brief instructions like, "Come up with a slogan for a hardware store." Instead, convey all the necessary information: "You're a copywriter, and your task is to create a slogan for a hardware store called such-and-such. Its key features are this and that. The director would like the slogan to reflect the aspects we discussed. Furthermore, it's important that the phrase be no more than five words long."—and so on. The more details you provide, the more accurate and relevant answers you can receive.

User Experience Manager, creator of the Telegram channel "Ask "Kira."

Imagine you've hired a new employee. They're enthusiastic and driven, but unfortunately, they lack the ability to anticipate your thoughts. It's the same with neural networks.

Imagine artificial intelligence as your assistant. Don't limit yourself to brief commands like "Come up with a slogan for a hardware store." Instead, provide it with a full description of the task: "You're a copywriter, and your task is to create a slogan for a hardware store called such-and-such. This store has its own unique features, such as such-and-such. The director wants the slogan to reflect certain ideas, such as this and that. It's also important that the phrase be no more than five words long," and so on. The more detailed you describe your requirements, the more accurate and relevant the answers you will receive.

Turn AI into your virtual advisor or assistant.

A marketing specialist at an international IT company and a mentor at the Higher School of Economics.

— I recommend using artificial intelligence to organize your thoughts and identify flaws in your idea or concept.

Use the duck method—it's an effective way to clarify your thoughts. You can voice your idea out loud or write it down, as if explaining it to a rubber ducky. In this process, you often identify errors and flaws in your logic.

When a lot of doubts and ideas arise in my head, I begin to voice them to the neural network. I simply lay out my thoughts, expressing uncertainty and pointing out points that raise questions. This process allows me to organize my thoughts, identify flaws in logic, and better understand what exactly I want to convey. If after this "duck dialogue" the problems still haven't been resolved, I turn to my colleagues for advice. However, by this point, the underlying difficulties are usually already clear.

I like to turn to AI to act as a critic. When I need to test the robustness of my strategy, I ask the neural network to be skeptical. She begins to express doubts about my plan: she points out vulnerabilities, raises possible risks that might have gone unnoticed, and helps me notice nuances that are easy to miss in a state of enthusiasm.

Marketing Specialist in a global IT company and a mentor at the Higher School of Economics.

— I recommend using AI to organize your thoughts and identify flaws in an idea or concept.

Use the ducky technique. This approach helps you better structure your thoughts. You verbalize or write down your idea as if you were explaining it to a rubber ducky. During this process, you often discover your own shortcomings and logical inconsistencies.

When a lot of doubts and ideas arise in my mind, I begin to share my thoughts with the neural network. I lay everything out as is: my thoughts, uncertainties, and points that seem unclear to me. This allows me to organize my thoughts, identify logical gaps, and better understand what exactly I want to convey. If after this "dialogue with a virtual assistant" there are still unresolved questions, I turn to colleagues for advice. However, more often than not, by then the problem has already become obvious.

When I need to test my strategy, I turn to AI to act as a critic. It begins to analyze my plan, identifying its vulnerabilities and raising questions about potential risks I might have overlooked. This approach allows me to see nuances that sometimes escape my attention when I'm too carried away by my idea.

Outsource routine tasks to a neural network to add interest to everyday activities

Project Management and Content Marketing Specialist at the Sdelayem Agency

Every day, I encounter a huge amount of information—facts, tasks, corrections, and notes. To keep my sanity in this deluge, I turn to neural networks. I primarily use the paid version of ChatGPT, but sometimes I experiment with DeepSeek. These tools can really make my work much easier.

Set reasonable deadlines for tasks. When a new project comes in, especially if it consists of multiple stages, like creating a landing page or a series of articles, I don't want to have to set timelines and calculate the costs for each element myself. Instead, I simply feed the neural network with contract information: the start date, deadline, and the number of days it takes designers, authors, and editors to complete each stage. The result is a kind of mini-assistant—a manager who doesn't get tired or forget about deadlines.

I try to format input data so that it's clear and easy to understand. When I need to calculate a KPI or engagement level, I don't waste time studying formulas. Instead, I simply list the project's inputs: what exactly needs to be calculated, what data I have, and what format the table needs to be in. The neural network suggests various metrics and instantly generates the corresponding formulas. I select the most suitable option, adjust it slightly to suit my needs, and the process is complete. This saves a lot of time, especially if you are not a professional analyst, but creating tables is still necessary.

Project manager and content marketing specialist at the Sdelayem agency.

Every day I I'm faced with a huge amount of information: facts, tasks, corrections, and notes. To keep my sanity in this flood, I resort to the help of neural networks. I mainly use the paid version of ChatGPT, but sometimes I experiment with DeepSeek. Here's how they can support me.

Set reasonable deadlines for tasks. When I receive a new project, especially if it involves many stages, such as creating a landing page or a series of articles, I don't want to set timelines and calculate when and what needs to be delivered. I simply feed the neural network with information from the contract: the start date, the deadline, and the required time for each stage performed by the designer, author, and editor. The result is a kind of mini-assistant—a manager who doesn't get tired and remembers deadlines.

I try to formulate the input data in a way that makes it as clear and convenient as possible. When I need to calculate KPIs or engagement levels, I don't waste time studying formulas. Instead, I simply describe the key parameters of the project: what indicators need to be calculated, what data we have, and what format the table should be in. The neural network suggests various metrics and immediately generates the corresponding formulas. I select the most suitable option, modify it slightly to suit my requirements, and the task is done. This saves a lot of time, especially if you are not a professional analyst, but the need to create tables is still there.

If you are interested in neural networks, we recommend checking out other materials from Skillbox Media.

Without causing awkwardness, but on the contrary, causing delight: how companies should develop advertising using artificial intelligence so that it is effective and not irritating.
The problem of the digital age: ways in which ill-wishers can use deepfakes to damage a brand's reputation.
Is GEO our only chance to maintain our position in the era of artificial intelligence? Let's discuss this with the professionals.
Marketing trends expected in 2025: how artificial intelligence, search engine optimization, video and influencer marketing will develop.
Mastering neural networks and leaving your usual five-day job: the story of one of the first artists using neurotechnology.

Practical study of neural networks

Find out more