Successful AI transformation starts with people.

Generative AI holds enormous economic potential. A Stanford and MIT study, for instance, found productivity gains of 15% in customer service.¹ Inside most companies the reality looks different: only a small group of around 5% are already scaling real results with AI, while the large majority record minimal financial return despite substantial investment.²

01 We talk too much about models and too little about people.

Investing in the technology alone isn’t enough. Germany’s recent past says as much: despite digitalisation, labour productivity has barely grown for years, and in 2023 it actually fell.³ Not because technology doesn’t work. Some areas raised their efficiency enormously through digitalisation, while elsewhere elaborate slide decks were built where a flipchart would have done. Too often things were digitalised for the sake of digitalisation, rather than for any real benefit.

Rolling AI out across the company as the next IT tool, with no clear idea of what it’s for, repeats exactly that mistake: the emails suddenly read better than ever, and the financial benefit never arrives, because the actual value creation was never touched. To realise that potential, AI has to be understood as a lever for rethinking processes, roles and ways of working from the ground up. That’s precisely what the Boston Consulting Group’s 10-20-70 rule quantifies: in a successful AI transformation only 10% of the effort goes on the models and 20% on data and IT infrastructure, but a full 70% on transforming business processes and the people who work with them.⁴ Because an interface nobody opens is worthless.