AI fucked up the world we live in.
And yet I keep using it every single day.
AI now accelerates app delivery, streamlines processes, and organizes schedules efficiently.
I kept hearing that all this technological boom is not real AI. “These are LLMs (Large Language Models) that predict what to answer next” – I hear.
Call them LLMs, call them autocomplete on steroids – the label does not change the fact that entire workflows have been rebuilt around them.
I don’t really care about the definition.
All I care about is the fact that from startups to giants, companies invest in AI and LLMs to enhance profitability.
I’ve witnessed AI’s success firsthand, contributing to impactful business implementations. The fact that all of these have been prioritized above other goals shows how important these projects were.
One of the obvious, yet, extremely useful usage was meeting reporting.
Many of you already know that coordinating meetings across time zones is a challenge in a global organization. It usually means chasing replies for days before anyone acted.
Once you finally get everyone on the same call, you need to introduce people to the topic, problem and figure out the solution.
It is very rare that you would resolve bigger matter on one call, so the task keeps repeating over and over again and each time someone skipped the meeting had to be realigned with what happened.
Once AI meeting reports had been implemented everyone was pretty aligned on everything right after the meeting and it was making the job way smoother. They could simply read the actionable report with points discussed and what is happening next with a person assigned to each specific, actionable task.
Before, there was usually someone who had to understand everything that was happening in this meeting and write it down in a way that was understood by everyone and make sure it was actionable enough, so the project is moving forward.
Making it actually work took a few weeks of prompt tuning, integration fixes, and convincing people to trust the output. Ultimately the tool works well and I believe that at this point people simply got used to new standard.
Here comes another familiar scenario for many.
Weekly reporting.
Oh lord, I remember working on each report on Friday to make it actionable, memorable, and short enough not to lose everyone.
AI had clearly improved this part, as you could simply write everything down together with dedicated internal AI tool and get rid of everything that was not bringing value to the leadership.
I can tell you that most of what teams consider progress is noise to the people above them, and AI turned out to be a surprisingly honest editor.
It was especially helpful if you had to analyze a big chunk of data and understand the trends and correlations between items. Something that could have been taking days or even weeks was completed within a couple of hours or days. Obviously, it was still necessary to review the data, but clearly it already was a big improvement.
Another scenario where I found AI to be extremely useful was online networking.
For about a year I was running outbound sales, which meant living inside LinkedIn and measuring my worth in reply rates.
The sales stint taught me the clearest lesson: AI is excellent at finding the right person and completely useless at making them care about you.
Don’t get me wrong, I still believe that the automation I built was extremely useful because it let me meet people I could not have found otherwise. The thing is that it became tricky to actually build connection with these people, so they won’t sense that AI does the entire job.
People can smell bullshit very quickly and once you burn your first impression then it is extremely difficult to actually get the relationship going.
What AI cannot do is make someone feel seen, and that gap is exactly where I had to show up myself.
None of these workflows ever made the work feel meaningful, but they gave me back the hours I needed to do the part that does.
Every workflow I automated bought me time, and I learned the hard way that the only thing worth spending that time on is the people AI cannot reach.