It’s a Partner, Not a Competitor: How Professionals at AdTech Holding Use ChatGPT in Their Work
OpenAI’s ChatGPT has sparked a lot of curiosity and debate about the future of various job positions. No wonder: this chatbot can not only support a conversation but also solve various technical tasks.
However, ChatGPT is unlikely to replace most jobs anytime soon. On the contrary, it has become a reliable partner for many IT and digital marketing professionals, helping them improve their work efficiency — not working instead of them!
We spoke to experts from AdTech Holding to find out how they harnessed ChatGPT to boost their work results.
ChatGPT as the new encyclopedia
‘Google it’ is rapidly changing to ‘Ask ChatGPT’ — and it’s not just about trivial questions and quick, simple requests. This makes sense: Google search implies analyzing dozens of links and skimming long reads to find information, while a chatbot gives a terse reply to an exact question.
Here is what Alex Pilov, the Team Lead of the AdTech Product Team, shared with us:
‘Most frequently, I ask ChatGPT about technologies I am not quite familiar with — for example, questions about a standard API of a programming language I don’t use.
Another example: I sometimes forget particular settings of docker-compose.yaml. It’s much easier to ask ChatGPT than to sift through StackOverflow or technical documentation. You get a quick reply with examples — and this saves much time.
ChatGPT as a junior coder and QA specialist
In fact, ChatGPT was not specifically designed to perform programming tasks. However, it soon turned out that it could definitely generate code snippets in popular languages, including Java, Python, C#, and more.
We got the first evidence of that from Peter Garmashov, the Head of Products —
‘ChatGPT wrote a landing page code for me yesterday’ — he shared.
Of course, we cannot reasonably expect that the bot might replace an experienced programmer in the nearest future. Still, it is capable of assisting with coding tasks in some rather creative ways.
Here is an example we got from our QA team: Mikhail Sidelnikov, the Head of QA in the Product Team, shared a case study from his practice.
‘We needed to track redirect chains for one of our tests— and it required collecting browser requests. This is a pretty standard task, but we don’t solve such problems often enough and thus have no examples of the required code.
So Anna Masitseva, our QA specialist, went to ChatGPT and simply asked it something like this:
‘Write me autotest in java using selenide to get browser requests for page google.com’.
And the chat sent back the reply! The first output was not exactly what we needed, but we made our request more specific and got a code that worked tolerably well. Then we finalized it manually — and implemented it in the end. I am very proud of Anna’s idea — and I suppose we have to keep engaging AI to help us with routine and creative tasks.’
The human input is obviously irreplaceable, and you cannot simply rely on ChatGPT as an outsourced programmer. As Alex Pilov put it, ‘I can experiment with generating, refactoring, or validating a code via ChatGPT. However, I rarely use it for my everyday tasks — they are non-trivial, and I still lack trust in the AI in this matter.’
Still, as you see, ChatGPT can at least indicate the right direction for solving coding problems.
ChatGPT as the personal assistant
Nikita Ableyev, the Senior Product Marketing Manager of AppLabz, works on promoting the apps to the market. This job involves a big range of tasks — from finding new approaches for ads and ad groups to generating copy and images for creatives.
Nikita shared how ChatGPT helps him automate part of his work — and gave some unexpected lifehacks that might become insights for everyone dealing with tables, images, and translations.
- Working with texts
When we develop an app, we need to promote it worldwide. Thus, creatives for our ads must be localized depending on the country we are targeting in the particular campaign. Of course, the Google Translate tool can do this work in a couple of seconds — but the trick is that we need a particular number of words in every text: they must fit in the banner.
With ChatGPT, this task became much simpler: I give it the texts I need to translate and mention that the message must contain no more than X words in every line. Here is what I get as a result — a convenient table, ready to copy and paste to the creative:
The best part of this story is that A/B tests show a better performance for banners with ChatGPT translations than ones with Google Translate work:
- Helping with marketing research
Promoting an app in various countries might involve tracking different information. For example, one of the tasks required combining groups of countries with approximately similar Consumer Price Index (CPI).
Instead of searching for information on every country from multiple sources, you can simply ask ChatGPT to create a ready analytical table like that:
- Performing miscellaneous tasks
Besides, ChatGPT can help with finding keywords for app descriptions, providing you with brief descriptions of YouTube video content via special plugins, creating working emails, and answering simple or complex questions.
Here is an example of how the Video Insights plugin works to save time on watching videos:
And here is how the Chat solves advertising budget problems:
ChatGPT as a funny colleague
AI can sometimes be creepily realistic — especially when it keeps up a conversation in a very human-like manner. So why not use it as a little distraction from work?
Alex Pilov shared that he sometimes asks the Chat to tell a joke — but, in fact, it is much better in coding than humor.
Final Words
As you can see, ChatGPT does not seem a threat to professionals — and, with the right approach, boosts performance, allowing one to focus on tasks that require 100% human involvement.
