Many entrepreneurs wonder whether AI is hype or here to stay. There is one industry where that debate was settled twenty years ago: exchange trading. Computers took over the routine work completely, and whoever did not adapt disappeared. I worked on the trading floor of IMC Trading for 3.5 years, and the habits from that world turn out to be exactly the habits that decide whether AI succeeds or strands in an ordinary business. These are the five most important ones.
Why the exchange is the proving ground of automation
Trading firms such as IMC Trading trade automatically on exchanges around the world, in fractions of a second. Machines place the orders, people design the strategies, guard the risks and step in on exceptions. That division of labor did not appear because it was fashionable, but because every millisecond and every basis point is measurably worth money. If automation gets tested mercilessly anywhere, it is there.
At IMC I delivered daily analysis as a macro analyst to more than 1,000 traders worldwide. My job was filtering information under time pressure: extracting what was decision-ready from a stream of news, data and opinions, before the moment passed. Since 2022 I have applied that same discipline to AI, first as a power user, then as a developer and AI coach.
Lesson 1: an advantage only exists once you measure it
In trading, an advantage is called an edge, and an edge you cannot measure does not exist. Every AI application in your business should pass the same test: how many hours does this save per week, which mistakes does it prevent, what does it deliver in euros? "It feels faster" does not count.
The practical translation: start every AI application with a baseline (how much time does this process cost today) and agree upfront when the experiment counts as a success. If an application does not clear the bar, stop it. An AI application without measurable gains is not an edge but a hobby.
Lesson 2: separating signal from noise is the real work
Traders drown in information: news, data, rumors, opinions. The scarce good is not information but filtering. In 2026, entrepreneurs are in exactly the same position with AI: new models, new tools and new promises every week. The question is not "what is new", but "what is decision-ready for my business".
My rule of thumb: do not follow the news, follow your own bottlenecks. List the three processes that cost you the most time this month, and judge every new tool exclusively against those three. Everything else is noise, however impressive the demo.
Lesson 3: think about what can go wrong first
A trader thinks about the downside of every position first: what do I lose if this goes wrong, and do I survive it? Only then comes the profit. In AI implementation that reflex is rare, and that is exactly why companies get burned: an agent giving wrong answers to customers, sensitive data disappearing into a free tool, a process that becomes fully dependent on a single vendor.
Ask three questions upfront for every AI application: what is the damage if the answer is wrong, who checks the output before it goes out, and can we go back if the tool disappears? Only automate fully once those answers are good. Until then, keep a human in the loop.
Lesson 4: machines do the volume, people do the exceptions
The exchange settled the division of labor between human and machine long ago: machines do the repetitive volume work fast and flawlessly, people do the exceptions, the oversight and the strategy. Nobody on a trading floor tries to beat the machine on speed, and nobody lets the machine set strategy unchecked.
That is exactly the model for your business. Let AI handle the draft emails, the summaries, the standard questions and the admin, and put your people on the cases that deviate: the angry customer, the odd request, the decision with real consequences. Whoever uses AI to replace people one-for-one is copying the wrong part of the model.
Lesson 5: be judged on results, not on stories
On a trading floor you are judged daily: your analysis was right or it was not, and it shows in the numbers. That culture forces honesty. The AI world is different: there is no scoreboard, so good stories survive much longer.
Bring the scoreboard yourself. Ask everyone selling you AI advice, me included, what will change measurably and when you will see it. In the comparison of AI coaches in the Netherlands you can read which checks help: does the coach build himself, does he also say no, and does he know your context?
What this means for your business
You do not have to rebuild your company into a trading firm. The lesson is the method, not the scale: measure every application, filter on your own bottlenecks, think about risk first, divide the work between machine and human, and demand results. With those five habits, AI is not a gamble but a controlled experiment.
This is also exactly how I work in the one-on-one AI coaching and training: no tool parade, but your three biggest bottlenecks, a measuring stick and a division of labor that holds. You can read more about my background on my profile page.
Frequently asked questions
What is high-frequency trading? High-frequency trading (HFT) is automated exchange trading in which computers place orders in fractions of a second based on pre-designed strategies. Trading firms and market makers such as IMC Trading use it, among other things, to continuously quote buy and sell prices, keeping trading on exchanges liquid.
What does high-frequency trading have to do with AI for businesses? HFT is proof at scale that automation works when you measure, manage risk and put people on the exceptions. The same principles decide whether AI succeeds in an SMB: measurable gains per application, a human in the loop where mistakes do damage, and machines on the volume work.
Why is a trading background valuable for an AI coach? Because the core habits transfer. Tarik Eraslan delivered macro analysis to more than 1,000 traders at IMC Trading for 3.5 years: filtering information under time pressure for people who made money with it directly. As an AI coach he applies the same discipline: separating signal from noise, only counting measurable advantages and looking at the risk first.
Does my business need to become as automated as a trading firm? No. The scale of HFT is irrelevant for almost every business, but the method is not: start with a baseline, automate the repetitive work, keep people on the exceptions and stop applications that do not clear their bar.
Getting started yourself
If you want to apply this approach to your business, start with the three processes that cost you the most time today and hold them against the five lessons above. If you get stuck or want to move faster, book a no-obligation call: half an hour in which you present your situation and hear honestly where the measurable gains are. Rather learn the basics yourself first? Have a look at the one-on-one AI training.
