MEDIUM  –  Jérôme Bouquet  –  3rd June 2019

Stephen Hawking said that “intelligence is the ability to adapt to change”. As human beings we are characterised by intelligence and the capacity to learn and to understand, attributes that have enabled us to adapt over the course of millennia. Today, we need to confront and adapt to the world’s complexity and the exponential transformation in exchanges and knowledge. Artificial intelligence (AI) and the development of intelligence-simulating algorithms allow us to grasp this complexity and up our pace. AI grants power in a market, battlefield or territory to anyone who can integrate it into their use cases, solutions and devices.

Connecting uses and technologies in a short time is a real challenge. Hackathons, startup scooting, ideation, design thinking, pitching and reverse pitching are just some of the methods adopted to match uses with technology. It might seem simple. But for these methods to produce concrete results, a range of conditions must be met, in particular in the minds of employees and middle management. Putting trust in collective intelligence, and experimenting with possible courses of action rather than managing risks, put us in conflict with established processes, organisation and culture, generating disquiet that a good intrapreneur needs to deal with. The challenges organised by the DGA, the French defence procurement and technology agency, provide good examples of methods that help make the leap from ideation to operational deployment in the organisation.

In the natural economic selection to which we must adapt, it is important to understand what AI is, not from a technical standpoint but from an economic one. The power of AI stems from its network effects. The more relevant data an AI incorporates, the better it will be. In other words, the more users it has, the more data it gathers; and the better it becomes, the better its knowledge of customers and threats, its predictions, and its performance will be. The better it is, the greater its appeal will be to users and customers. As this product snowballs, it becomes by far the best way of serving users, and therefore a crucial feature of the landscape.

Anybody developing a tool or service using AI therefore gains a real edge in being the first to bring that product to market. In a Darwinian adaptation process, our survival depends on speed. Speed is of the essence if we are not to be elbowed out of the market or become subordinate to somebody else’s AI.

Survival depends less on technological excellence than the speed with which we deploy AI.

The Americans have made no mistake by reorienting strategically en masse in this area. Google CEO Sundar Pichai has announced that his company is making AI its priority strategic focus. Facebook is not to be outdone, bringing augmented reality together with AI. Elon Musk set up the non-profit OpenAI to share AI research and combat the privatisation of intelligence.

In France we have excellent engineers, mathematicians and IT specialists and are renowned worldwide for the excellence of our training courses. However, we have been overtaken, eclipsed somewhat by the US when it comes to internet technologies, and the same thing is happening with AI technologies. Does this mean France is too slow to release innovative products? Probably, yes.

In a challenge organised by the DGA in early 2017, the armed forces gave French engineers one year to come up with a product they could trial to meet the urgent need to adapt to a constantly and rapidly changing threat. The discussions around the definition of the challenge in the project team were quite representative of the relationship to time in French culture. While users stressed the urgency of the need and accepted that it might only be 60% met, engineers proposed a perfect solution meeting 100% of needs and complying with all standards, for delivery in… 3 years. Which to choose? Perfection in 3 years or a partial solution now? In France, there is often a preference for the former, while the laws that govern our environment and network effects have a natural preference for the latter. Similarly, users accepted an open statement of the challenge in the form of a story based on their day-to-day, while the engineers wanted a very precise statement of the need, even a solution, which would have shut down innovation. The team responsible for the challenge was able to withstand pressure to extend the lead time and focus on the short-term basics.

Geert Hofstede’s cultural dimensions theory shows that, unlike in Anglo-Saxon cultures, aversion to uncertainty is a characteristic of French culture, as it is of German culture. This is reflected by a large quantity of rules, norms, a need for precise planning upstream of any project, a refusal to allow grey areas and ambiguous situations, and close oversight by experts. Consequently a steep barrier is placed in the path of rapid innovation, a barrier that an intrapreneur needs to manage: to innovate you have to explore uncharted avenues, and to leave room for new ideas to grow by allowing vagueness and ambiguity to remain. French culture asks project sponsors to specify the timetable, risks and ROI in advance, closing down options and opportunities for innovation. The Anglo-Saxons, the inventors of agile methods, are more trusting and encourage rapid user-facing experimentation. This allows them to learn a lot about the need and the real benefits of technologies, which is impossible with expert analysis, however in-depth it is.

Speed in increased by accepting uncertainty and experimenting quickly.

Artificial intelligence helps us, based on data, to learn, understand and make decisions. It poses questions for our collective intelligence: how do we learn? How do we share information and knowledge? How do we gather data and build an understanding of it? How do we coordinate our individual intelligences to make decisions?

As Geert Hofstede shows, in the top-down organisational model preferred in French culture, decisions are made by the central authority. Data is used to justify decisions after the fact and to monitor lower-ranking teams. Our mental software tells us, consciously or unconsciously, that to disclose our data is to accept oversight and intrusion. Yet to learn collectively and to teach an artificial intelligence, we need to share data, ensure information is transparent and place great trust in the collective. For these reasons, the approach adopted by management crucial: less control, more autonomy, more trust. Managers need to move from an approach is based on control and decision-making to one based on delegation and collective intelligence. They must focus more on vision and allow operational teams to make decisions collectively. The Anglo-Saxons, having more of a negotiation culture, are better prepared to share and work cross-functionally. The Americans flip the pyramid. Steve Jobs once said: “It doesn’t make sense to hire smart people and tell them what to do; we hire smart people so they can tell us what to do.”

The innovation needed to identify use cases does not run on authority. You can order somebody to comply with a procedure or a decision; you can’t order her to innovate, to give away her ideas and to commit to creating something new. In other words, innovation is first and foremost a personal choice that requires autonomy. This autonomy is as much the fruit of freedom granted by the management as it is a form of personal and collective emancipation. The things that motivate us to innovate are very different from those that motivate us to fill a role in an organisation. They depend on a need to achieve and push one’s limits for a cause or a vision rather than on a need for self-esteem and material gain. Here too, the US has figured out these drivers, adopting very horizontal organisational structures comprising small teams with new coordination mechanisms.

Wiring autonomy into teams’ mental software is a way of developing AI-assisted collective intelligence.

To create the DGA’s programme of challenges, we encourage teams on the ground to identify use cases by extending their freedom of choice and, most importantly, by making the challenges meaningful for the Ministry for the Armed Forces and for each stakeholder. Contrary to a command-and-control system driven by each person’s individual interests in their post (responsibility, title, promotion, bonuses), innovation projects are driven by a goal that goes beyond the individual.

Artificial intelligence allows us to grasp the complexity of the modern world. To capitalise on it fully and survive in the face of global competition, we need to adapt our collective intelligence and move beyond our traditional operating methods. This cultural transformation is hard and time-consuming. Our culture is a bit like layers of sedimentary rock that have built up over time, over the course of our long history and that of the company. Against this backdrop, we need to rewire our collective mental software to help us adapt. Sébastien Bazin, CEO of Accor, put it well: “If we don’t figure out that the digital transformation is a culture change, we’ll never make it.” Standard change methods cannot work because this transformation cannot be planned. It can only be effected through action and by tackling issues relating to speed. Far too many people trust only in management committees, because they think adaptation needs to happen here. It is on the ground in employees’ mental software that it will all be played out.

To harness the power of AI we need to rewire our collective mental software.

In any company, cultural change takes a long time. However, it can be done locally and in a continuous series of small steps. There are three ingredients for projects that seek to transform the system.

The first ingredient is the completion of short-term practical projects that call an organisation’s modus operandi into question. Two one-year challenges set by the DGA, “Acquire an indoor drone with disruptive innovation” and “Acquire an AI solution to aid the interpretation of satellite images”, meet key Ministry for the Armed Forces needs that touch on current issues. Acquiring solutions in less than one year is a real challenge for the Ministry, which is accustomed to long decision-making and procurement processes and three-year lead times for the deployment of equipment in the armed forces.

Numerous organisations are setting challenges, hackathons and projects that prove uninspiring because of the barriers inherent in the way they are formulated. There can be no innovation or collective intelligence in a project where everything is fixed, the risks are already eliminated and the path is fully plotted.

The Defense Advanced Research Projects Agency (DARPA), a US defence department agency, chooses its projects based on a high level of scientific ambition. Elon Musk sets the level of ambition of his projects according to the limitations of physics. He starts from a long-range vision, checks it complies with the laws of physics, and translates it into short-term goals. These ambitions force everyone to think outside the box, to find new solutions, and to explore new coordination mechanisms. We too have to choose projects according not to the limitations of our organisations, but to the limits of the laws that govern our environment. In another more flexible and dynamic environment, what project could be accomplished that offers more added value for the organisation? This project must represent a long-range vision that can be completed in the short term. In France, with our very top-down, risk-averse culture, people too often tend to determine the budget, the scope and the level of risk, and to make lead time an adjustment variable. That’s why we often come to market too late or remain on an incremental innovation path. To bring about change, this equation needs to be flipped: the lead time should be decided first, so that the solution is made available to users within this short timeframe.

We need to choose projects that push the boundaries of our organisations and culture.

When we complete concrete projects that push the organisation’s boundaries and its coordination mechanisms, we produce precedents that can be communicated within the organisation, creating new operating models that can be disseminated. In practical terms, this helps refresh the organisation’s mental software. When it comes to the Ministry for the Armed Forces, moving from a long lead time to a short one helps to rearrange the equation. We want to aim for a score of “12 out of 20 now” rather than “20 out of 20 in 3 years”.

The second ingredient is change in the management’s approach. The management has to create the conditions for autonomy by having an inspiring vision and challenging the organisation; together these will guide decision-making within operational units and cross-functionally. The management must thus become more of a benevolent coach, a facilitator for projects that help achieve the vision. The DGA has identified the need for open, short-cycle innovation based on civilian technologies. The management has backed exploratory initiatives that put this vision into practice, for example by supporting the first indoor drone challenge and its transformation into an innovative procurement process. It has supported the closer relationship with the armed forces, the creation of a short decision-making loop, and the transformation of projects arising from challenges into concrete operations.

When management turns coach and delegates leadership to the field, it instils the autonomy needed to rise to the challenge.

The third ingredient is the leadership of the innovation project. This is without doubt the most delicate aspect and the hardest role to play. This role is a pivot between the various entities involved on the ground, between the established culture and the new culture to be shaped, between the management and teams, and between the different intelligences. Unlike a conventional project management role focusing on planning and risk management, this role incorporates more situational and emotional intelligence. Like gardeners, leaders must be able to create the conditions for collective intelligence, seize opportunities, fertilise ideas, start small while also growing the project, and protect the team’s culture of autonomy.

A challenging short-term project, a management team with a facilitating approach, and on-the-ground leadership are the three ingredients that have cultivated innovative projects in the Ministry for the Armed Forces. They are factors that help to change the collective mental software so that our technologies can be learnt and taught.

In short, not adapting rapidly poses a real threat. Numerous speeches stress that our country risks being overtaken and impoverished because of our lack of speed. And yet, we have exceptional talents that we struggle to build upon because of our aversion to uncertainty and our taste for power. Many approaches and methods are being promoted to innovate and integrate AI. But if we fail to factor in the cultural aspect, these approaches will remain artificial and deliver no real added value or real change.

A new horizon is opening up to us and we have all the talent and ability we need to face this unknown. Rather than programming artificial intelligences, we need to reprogramme our collective intelligence through concrete and challenging projects, and embrace the opportunity of uncertainty. Rather than controlling, we need to enhance teams’ autonomy through an inspiring vision. Rather than planning, we need to test and learn, experiment and observe.

In the words of Andrew Hunt, one of the authors of the Agile Manifesto: “rather than construction, software is more like gardening”. To programme an artificial intelligence, we need to cultivate our collective intelligence.

Share This