How to Prepare Your Business for the AI Revolution: Key Insights from Working with a Game-Changing AI Application

In today’s digital world, AI and machine learning (ML) are dramatically transforming the advertising landscape, especially in social media. Platforms like TikTok, Instagram, and YouTube Shorts have leveraged sophisticated ML algorithms to create personalised, attention-grabbing experiences that drive billions in advertising revenue. While the use of AI to influence user behaviour raises ethical questions, its effectiveness is undeniable. 

To dive deeper into this topic, we’re speaking with Vitalii Chukhlantcev, an expert in advertising technology and analytics, who has experience working at one of the fastest-growing AdTech ML startups that directly competes with social media giants for advertising budgets. Vitalii has dedicated much of his career to understanding how to maximise AI’s impact in the AdTech space and believes the principles of successful AI applications in advertising can be applied across various industries. In this interview, he will share insights into how AI is revolutionising advertising and what businesses can learn from the advertising industry’s experiences with machine learning.

 

What Key Factors Should Businesses Consider When Deciding Whether To Implement AI Technologies In Their Operations?

 

When choosing an AI solution, there are five key factors to keep in mind: short-term ROI, long-term ROI, scalability, reliability, and simplicity.

The perfect AI solution really comes down to having a positive answer to all five of these considerations. That said, a solution can still be good even if it only meets some of the criteria — though ideally, you’ll want all five boxes checked.

First and foremost, you need to be sure that AI will deliver a positive return on investment in the short term. It’s essential to pick a solution that is cost-efficient and will actually save you money or generate revenue right away. If you’re not seeing a financial benefit soon after implementation, it might not be the right fit.

The long-term profitability of the project is even more important than short-term ROI, but with the rapid pace of technological change, it’s nearly impossible to predict what the best solution will be for your needs two years from now. Personally, I like to estimate the long-term ROI, but I always focus more on the immediate impacts a solution can offer.

Next up is scalability. A great AI tool should scale effortlessly with your business, whether you have 10 clients or 1,000. As your company grows, the solution should be able to grow with you — not require you to rebuild your entire operations process from the ground up.

Reliability is another critical factor, and it’s one of the biggest concerns in the industry when it comes to AI. If you can’t trust AI to consistently respond to customer queries or handle tasks correctly, then it’s not worth integrating. AI still has its issues — it can “hallucinate” or make errors, especially with multi step tasks. That’s why it’s important to test it thoroughly and use it for tasks where there’s some room for error, like creative work or idea generation.

Finally, simplicity is key. This doesn’t mean you shouldn’t push the boundaries of innovation when implementing AI, but the problem you’re trying to solve should be narrow and well-defined. As any tool, AI works best when there are clear guidelines and expectations in place.

 

Once A Decision Has Been Made, How Can A Business Assess Its Readiness For AI Implementation, And What Resources Are Typically Required?

 

The great thing about today’s AI tools is that the barrier to entry is lower than ever. When you look at the state of AI in 2024 compared to just three years ago, it’s clear that many of the complex algorithms we once relied on can now be used much more easily and creatively.

Take quality control in factories, for example. In the past, specific machine learning solutions had to be custom-built for each individual business. But now, with multimodal large language models like Gemini, you can simply write a small amount of code and ask the AI to analyse the quality of your products — using just words.

What’s even more impressive is that you don’t always need a lot of additional resources or preparation to get started. Essentially, all you need is the right data — whether that’s images, video, or text — and a simple script that connects that data to the AI via an API. And just like that, you’re good to go.

Once you’ve confirmed that you can gather the right data at scale and have tested it in a few trial runs, you’re pretty much ready to build a tech flow that feeds your data directly into the model. Since the main modality of interaction with AI is now natural language rather than code, deploying AI is a lot more accessible than it ever was.

 

 

As Companies Look To The Future, What Promising Applications Of AI Should They Be Aware Of?

 

First up, data-driven decision-making. AI is now capable of analysing huge amounts of data, helping businesses gain deeper insights into customer behaviour and fine-tune their operations. It’s especially useful for collecting data at scale that was previously hard to access.

Then, there’s enhanced search capabilities. Tools like Perplexity and Glean are already helping businesses find information faster, both internally and externally. And as AI-powered search gets better, it’s only going to become more reliable and efficient.

Next, we have automated content production. AI is already being used to create tons of content, especially in formats like reels. Generating marketing materials—whether it’s images, audio, or video—is not only cost-effective but also surprisingly easy to do today.

I believe these three areas will continue to grow and evolve. They also line up well with the five key criteria for AI integration, making them great starting points for businesses looking to dive into AI now.

 

Image: Vitalii Chukhlantcev

When Exploring These Applications, What Common Challenges Might Arise During The Implementation Of Scalable AI Solutions, And How Can Organisations Effectively Overcome Them?

One of the most common concerns with adopting new technology is the fear of starting from scratch. It can be tricky to figure out exactly where and how to apply AI, even if you have some initial ideas. A good approach here is to encourage a few trusted team members to get familiar with the available tools and test them out before deciding on the next steps as a group.

Now, going back to those three AI applications I mentioned earlier, I think it would be really valuable for members of the marketing, analytics, and IT teams to start using these tools in their everyday work. This hands-on experience can help generate a solid understanding of what’s possible and turn initial enthusiasm into practical, actionable ideas much faster.

Another big challenge is the technical ability of your team to integrate new systems into your existing workflows. You might not have the clean, standardised data needed to fully leverage AI, and that could be a roadblock. If that’s the case, it’s something you’ll need to address before diving into new analytical tools, for example.

Collecting, storing, and managing data as your organisation grows and changes is fundamental to scaling AI solutions. This could be a big (or small, depending on your perspective) bottleneck in driving innovation. The key is to ensure your database and tech teams are agile and that you’ve built a sustainable, flexible data infrastructure that can support future AI initiatives.

 

Can You Outline A Step-by-step Framework That Organisations Can Use To Successfully Implement Scalable AI Solutions?

 

First, start by defining the priority problem in your organisation. For example, let’s say it’s quality assurance.

Next, select team members who are motivated to solve this problem. Give them the time and resources to test the AI tool and brainstorm how to integrate it effectively. If testing isn’t feasible, skip this step.

It’s important to pay close attention to your data collection process. Good data is well-standardised, comprehensive, and free from human error, so you can trust the AI to handle it correctly every time.

Before fully rolling out the AI, run a few simple trial runs to make sure the results meet your expectations. Remember, unless the AI is fed the right data with the right prompts, the output will likely be poor at first. The goal at this stage is to perfect the flow until the AI produces the right results consistently.

Then, focus on building the IT infrastructure that will allow you to scale the approach. I’d also recommend setting up random output checks at this stage — these can be AI-driven too. If something breaks, you’ll get notified right away, and the issue won’t go unnoticed for months.

As long as your data collection process is solid and you’ve proven that the AI can handle this data well on a small scale, you’ll be able to scale the process effectively.

 

Finally, What Best Practices Should Organisations Follow To Prepare For Their First AI Solution Implementation, Particularly Regarding Team Training And The Necessary Technology Infrastructure?

Team training can be broken down into three key areas:

First, training project managers across different departments is essential. These individuals need to familiarise themselves with the tools and understand how they integrate into the larger system. The focus here should be on providing them with access to AI tools, encouraging exploration, and ensuring they have ample time to dive into the possibilities.

Next, training your technical teams may require more specialised instruction, although some team members might be able to pick things up on their own. It’s crucial that your tech team understands how your specific data infrastructure needs to evolve to support scalable AI applications. They also need to have the expertise to make those changes in-house.

Lastly, other stakeholders, such as those who will interact with the new tools, need to grasp the nuances of working with AI-generated content, including understanding any potential limitations or risks.

Update as of autumn 2024: The Telegram channel co-founded by Vitalii Chukhlantcev and his partners has officially launched and has already surpassed 100k subscribers. All interested individuals are welcome to join here (currently available in Russian language). The channel focuses on AI applications for both business and personal use, bridging cutting-edge technology with everyday life, while testing new tools, sharing impactful news, and summarising expert insights on the future of AI. It has been ranked in the top 10 for subscriber growth in the technology category by TG Stats.