A successful AI strategy requires an experimental mindset
Although artificial intelligence has been used since the 1950s, the technology has only gained in fame in the last decade as it is growing more and more out of its infancy. But when AI will actually be mature, it remains to be guessed for the time being. That makes it challenging to draw up an adequate AI strategy.
AI is more than ever a theme that requires attention in the boardroom. Many organizations see the potential of the technology, build small prototypes and develop use cases. Reinier van Leuken, director of product management at Salesforce and expert in the field of AI, finds it impressive that the business community has acquired a lot of knowledge on the subject in a relatively short time. ‘For me, this confirms that we can’t speak of a hype. The adoption curve is so steep that AI is already past the top of Gartner’s hypecycle.’
Large-scale adoption
This also agrees Jeroen van der Velden, associate professor of Strategy and Transformation at Nyenrode Business University. “I think we’re just past the top. From a strategic perspective, in the first phase, expectations and attention for a new technological development are extremely high, but particularly dominates the technological potential. The technological potential of AI is enormous, but we are now at the point where the technology needs to be integrated into an organization and there has to deal with a complex ecosystem of legal, ethical, social and organizational aspects.’ Companies successfully experiment with different forms and applications of AI, but get stuck when the Proof of Concept needs to be rolled out on a large scale. ‘For the large-scale adoption of business use cases, organizations still have some barriers to break down in this area,’ says Van Leuken.
Not to be predicted
The fact that AI will have a huge impact on our economy, society and business is as a stake. The crux is that we can’t predict how the technology and possible applications will develop. And that makes formulating an AI- strategy for organizations a big challenge. Van der Velden was involved in the rise of the internet in the eighties. “We already realized at the time that the internet would have a huge impact, but we had no idea what it would be. Many discussions at the time were about a paperless society, because the technology of the time, HTML, referred to a huge document with hyperlinks. So the idea was that you no longer had to print that, but could read it on a screen. That’s how far people’s imagination reached. It shows that we understand very well afterwards how a technology has developed and acquired an important position, but that we cannot predict that ourselves at the moment. With AI we are on the eve of such a same development.’
To find out what you want and can do with AI, you need to set up pilots with users. With your feet in the clay
Reinier van Leuken
Director of product management at Salesforce and expert in AI
Agile strategy
It is therefore crucial that an AI- strategy is agile and that space is created for development and a learning process. The organizations that can continue to observe, learn and adjust can follow the developments of AI. Those developments are still going so fast that it is impossible to predict how an organization should use the technology in five years. ‘It is therefore important to dare to experiment,’ says Van Leuken. ‘Certainly with a technology whose endpoint is not clear, fail fast is a formidable strategy to determine what works for your organization and what doesn’t.’ He stresses that AI experiments should not be done in a laboratory environment, but should be integrated into daily practice. ‘Organizations have to get rid of Proof of Concepts, because the concept has already proven itself. To actually find out what you want and can do with AI, you have to set up pilots with users, with your feet in the clay. Dare to stit your head, learn and thus develop your AI- strategy.’
Improve business process
Van Leuken sees two ways to start with this: improving an existing business process or looking for something the organization has never done before. ‘That choice is largely a matter of business objectives. If you look at the first, it is important to look at your own process with a fresh look and look for inefficiencies or improvements through the use of this technology. AI mainly brings that: efficiency, productivity and consistency. So you have to look for where processes sometimes run very well, but sometimes suddenly completely derailed and what would happen if those out-de-bocht-vliegers can be prevented. It often starts with ‘what if’ questions and out-of-the-box brainstorming about existing processes.’
If you keep making a better candle, you don’t suddenly get a light bulb
Reinier van Leuken
Director of product management at Salesforce and expert in AI
New innovations
The other start requires even further thinking outside the calibrated frameworks, where organizations do not look at what they can improve, but how things may be completely different or new. ‘If you keep making a better candle, you don’t suddenly get a light bulb,’ says Van Leuken. This also applies to the use of AI. ‘It requires board members to think experimentally. Take the self-scan at the supermarket. Someone at some point thought that we could provide the cashiers with enormous extra impact.’
The board must be able to view AI from all perspectives and monitor that these elements are included in decision-making
Jeroen van der Velden
Associate professor of Strategy and Transformation at Nyenrode Business University
Role for the board
That uncertainty about new possibilities ensures that organizations must be able to continuously adjust their strategy. AI is not only technology, but also brings political, social, ethical, legal, economic and other issues. According to Van der Velden, the board has a major role to play there. ‘They need to be able to view AI from all those perspectives and ensure that these elements are included in decision-making.’
Awareness and knowledge development
The question is whether many board members are sufficiently capable of this at the moment. ‘There is still so much uncertainty about AI,’ says Van der Velden. ‘Even the experts in this field don’t always agree on what they mean by AI. Many board members, I think, have a limited idea of what AI involves. I suspect they immediately think of ChatGPT, but it’s so much more comprehensive. That makes it difficult to have clues to make actual decision-making. That is why it is important that they create an environment in the organization where at least awareness and knowledge development takes place, so that they can steer it.’ According to him, an AI- strategy starts more from the perspective of awareness and exploration and creating space for it. ‘It is actually very clear that AI is not yet in the field of exploitation, but in the exploring part. So offer space for experiments and learning within your organization. Look for possible innovations. Actually, AI is an innovation process.’
AI requires an iterative process of learning, adjusting and researching. And then you come to a place you could never have predicted before
Jeroen van der Velden
Associate professor of Strategy and Transformation at Nyenrode Business University
Trust is core value
Both Van der Velden and Van Leuken are convinced that AI will have far-reaching consequences for organizations in the future. ‘AI will touch the heart of your organization,’ predicts Van der Velden. ‘It has an impact on our way of working, thinking, structuring information, not to mention the legal side. The GDPR will also play a role in this. It is good to have knowledge in your organization about how this works and what consequences this may have for your business model.’
Finally, Van der Velden urges board members to the heart of AI not to see as a one-dimensional, technological issue, but to take all the different aspects it affects. ‘A pitfall in strategy determination is too quickly delimiting your goal in this phase. AI is currently still so wide, the possibilities are still under development. If you already formulate too much of an endpoint, you are not going to store it up. AI requires an iterative process of learning, adjusting and researching. And then you end up in a place you could never have predicted in advance.’
Salesforce Einstein
Salesforce Einstein is the first comprehensive AI for CRM. It is an integrated set of AI technologies that make Salesforce Customer 360 work smarter and give organizations worldwide access to AI. Einstein enables organizations to embrace AI-led strategies, increasing in-depth insight into customer behavior and proactively anticipating needs.