The Anonymous Case Studies

I’ve been reading the same case studies for months and I can’t stand it anymore because it simply isn’t true. A mid-sized company deploys AI in customer service and six months later customer satisfaction is up by 35 percent. Another uses AI for workforce planning, efficiency rises by 40 percent. A third doubles its conversion rate with AI in marketing.

Sounds brilliant. But not a single one of these examples has a company name. No industry beyond a generic category. No contact person, no source, nothing you could verify.

I’m not someone who collects these examples and meticulously fact-checks them. But it stands out way past my threshold for tolerance. It’s everywhere and always follows the same pattern. An AI agency immediately identifies the gap, implements AI, and within three to six months the numbers improve by 25 to 60 percent. No complications, no resistance, no problems where something didn’t work. Apparently there was never an employee who had concerns or one of the notoriously overloaded IT departments and no costs that spiraled out of control either.

Sorry, but that’s nonsense.

I’ve been involved in many implementations throughout my career. Whether new technologies, logistics processes or sales strategies, not a single one went as smoothly as these examples suggest or as we’re being told. In reality there’s always a moment where something doesn’t go as planned. Data is never clean enough to migrate anywhere without problems, interfaces transfer everything possible except the desired data sets, employees don’t get on board, costs are higher than planned. In the AI examples mentioned, apparently nobody stands there afterward and says, that was a bad idea. That moment is missing from every single one of these examples. And I’ve had it often.

Either they only talked to companies where everything happened to go perfectly or the examples were tailored to fit any success story. I think a third option is most likely. It’s all shamelessly made up.

Anonymous examples aren’t inherently bad. Sometimes companies don’t want to be named. But when every example is anonymous, the situation is clear to me. The claims all follow the same arc. Before, everything was slow, manual and error-prone. After, everything is fast, automated and efficient. The improvement numbers always land in the same corridor. Never below 20 percent, that would be unspectacular. And above 70 would be unbelievable. High enough to impress, low enough to still sound plausible.

It’s honestly annoying. When this hype leads to spreading numbers like these, something is off. I don’t just wonder who produced these numbers but why. I could answer cynically and say it was an AI consultant who urgently needs new projects because he’s afraid of being rationalized away by AI himself. I read these examples like this: if they contain something that went wrong, they come from practice. If not, it’s bad salesmanship. Because AI is not the first product in the history of technology that works perfectly on the first try and produces miracle numbers like these.