In this Rise of AI Virtual Chat from September 2020, host Fabian Westerheide talks with Marcel 'Otto' Yon, who spent his last four years as founder and CEO of the German Armed Forces' digital innovation unit and earlier built companies in facial recognition, later used in German passports, and in image recognition for cancer detection. Drawing on roughly 50 corporate and military innovation units he has studied, Yon lays out five traits of the successful ones: a focus on execution rather than ideas, direct reporting to the CEO, positioning as an internal competitor to the core business, room to grow instead of an early handoff, and mixed teams of insiders and outside entrepreneurs. The conversation covers why German organizations move slowly compared with the US, citing accountability culture and business-to-government sales cycles of around ten years, why diversity in leadership means more than gender, how his unit built trust with soldiers, startups and allied militaries, the personal backlash he faced, and his view that AI's exponential growth will leave organizations too little time to adapt once the curve turns steep.
Key takeaways
- Yon says the most successful innovation units report directly to the CEO and act as an internal competitor to the core business.
- He argues large organizations must shift from 'critical mass' to 'critical initiative' since size now slows them down.
- He says his military innovation unit reached 40 percent women, showing gender diversity is achievable even in conservative settings.
- He attributes Germany's roughly ten-year B2G technology sales cycle partly to a culture that rewards avoiding mistakes over achieving results.
- He warns exponential AI adoption can look flat for years before crossing an acceptance threshold too quickly for large organizations to adapt.
Detailed notes
Five traits of successful innovation units
Yon says that after analyzing military innovation units across the Western world and about 50 corporate ones, five recurring themes separate the successful from the struggling: the unit's value lies in execution rather than ideation; the most successful units report directly to the CEO rather than to a CIO, CFO or lower management; disruptive units work best when positioned as an internal competitor to the core business rather than in harmony with it; projects should be allowed to grow into new core activities instead of being handed off early, since early handoffs tend to fail; and the strongest teams mix people from inside the organization with experienced outside serial entrepreneurs. He illustrates the reporting point with Walmart, which failed at digital transformation three times before succeeding on a fourth attempt once the CEO appointed a co-CEO dedicated to digital, and with the US Air Force, where a digital-innovation leader was promoted to deputy CIO of an organization larger than the entire German military.
From critical mass to critical initiative
Yon argues that large organizations need a cultural shift from what used to be an advantage, critical mass, to what he calls critical initiative, since size now makes organizations slower rather than stronger. He says most C-suites reflect the business as it has been rather than as it needs to be in ten years, and calls for more entrepreneurial diversity in leadership rather than boards composed only of people who know how to run the existing business. As an example of rapid cultural change, he cites Singapore's military and defense procurement organization, which moved a culture program from a small pilot area into finance, legal and other core functions and showed organization-wide effects within roughly four to five years. Westerheide adds his own experience advising a roughly one-billion-euro German family business, telling its leadership that Alibaba offers a comparable solution for free and could displace them within five years without their noticing.
Diversity beyond gender
Yon says his own military innovation unit reached 40 percent women, showing gender diversity is achievable even in a conservative environment, but argues diversity must extend further: to career background, since a startup founder who scaled a company from nothing brings broader functional experience than a corporate manager who has only ever run one function at increasing scale, and to psychological makeup, balancing cautious risk-managers with pioneers willing to step into the unknown. He notes that shareholder-appointed board members show little diversity in Germany, and argues that more diverse boards would cascade into more diverse C-suites and, in turn, into the wider organization.
Why the US and UK move faster
Comparing examples such as JPMorgan hiring machine learning experts at scale after the financial crisis against a struggling Deutsche Bank, Yon attributes the gap partly to accountability culture: American executives are held accountable for moving things forward, while German culture tends to hold people accountable mainly for not making mistakes. He describes business-to-business purchasing decisions taking two to three years in Germany versus six to twelve months in the US, and business-to-government decisions taking around ten years in Germany, citing his own facial-recognition company's decade-long wait for significant government revenue despite strong early interest after September 11. He also contrasts Germany's tendency to subsidize AI research and technology directly with US and UK approaches that fund demand for results, such as subsidizing police departments to adopt new technology, arguing this pulls solutions toward real use rather than pushing untested technology.
Building the German Armed Forces innovation unit
Yon describes building a team of officers and soldiers, including people from cyber special-forces backgrounds, alongside serial entrepreneurs from the startup world, and says every one of roughly 80 projects the unit initiated directly involved troops at the operating level. He says startups, contrary to expectations, were highly willing to work with the military once they understood its scope extended to healthcare, logistics and administration, not only weapons, and that partnerships with allied militaries' innovation units in the US, France, the Netherlands, Singapore and Israel were marked by openness despite slow official diplomatic channels. He credits a deliberate change-communication strategy, including a bottom-up community of change-makers that grew to around 10,000 members, with employee representatives becoming among the movement's strongest supporters rather than resistors.
Backlash and the future of AI adoption
Yon says pushing disruptive change in the German military brought anonymous hostile letters and personal, even family-directed attacks, which he attributes to a small number of people he calls saboteurs, criticizing the public sector's inability to remove them once identified; he cites a prior guest, Chris Boos, as suffering similar public backlash after funding a Corona app personally. Looking at AI's trajectory, Yon says the underlying technology, capital availability and mainstream acceptance of startups have all changed since earlier AI cycles, and warns that exponential technologies are especially dangerous for large organizations because the curve can look flat for years before crossing an acceptance threshold within one or two years, too fast for slow organizations to adapt. He expects the next five years to be pivotal, says AI adoption is held back mainly by a lack of innovation culture rather than by the technology itself, and predicts German corporations will either adopt a more entrepreneurial culture and lead, or be overtaken by new startups.
Quotes
Being an entrepreneur is like being in a boxing fight with endless rounds, and all you can do is take the beating and get up again, because there is no way of giving up.
Marcel "Otto" Yon
In the old days critical mass was an advantage in itself. Today it's a disadvantage because it makes me slower, so I need to move my organization to critical initiative.
Marcel "Otto" Yon
Alibaba is running, by the way, the same solution, is offering for free, and they can crush you in five years and you won't even see it coming.
Fabian Westerheide
Questions answered in this talk
What makes a corporate or military innovation unit successful?
According to Marcel "Otto" Yon at Rise of AI in 2020, drawing on around 50 corporate and military innovation units, the most successful ones focus on execution over ideation, report directly to the CEO, act as an internal competitor to the core business, are allowed to grow rather than being handed off early, and combine insiders with outside serial entrepreneurs.
Why is business-to-government technology procurement so much slower in Germany than in the US?
According to Marcel "Otto" Yon at Rise of AI in 2020, B2G purchasing decisions in Germany take around ten years compared with six to twelve months in the US, which he attributes to a public-sector culture that holds people accountable for avoiding mistakes rather than for achieving results, and to splitting decision-making authority from responsibility for outcomes.
How did the German Armed Forces build its digital innovation unit?
According to Marcel "Otto" Yon at Rise of AI in 2020, as its founder and CEO, the unit succeeded by mixing soldiers, cyber-special-forces personnel and serial entrepreneurs, involving troops directly in all roughly 80 projects, partnering openly with allied militaries' innovation units, and building a bottom-up community of change-makers that grew to about 10,000 members.