It’s the conversation that’s dominating board rooms and strategy meetings. How are firms preparing for, or already harnessing, the potential of AI for their organisation?
A KPMG survey found that 65 percent of executives anticipate a significant impact from generative AI (GenAI) within the next few years, even as many firms are still in the early stages of implementation. My own survey of INSEAD alumni found that 68 percent of business leaders were already using GenAI at work.
This has fuelled significant investments in the sector, with funding reaching US$21 billion last year. But there is still a distinct lack of certainty about what exactly GenAI’s impact will look like. Reports veer from wild optimism about the technology’s capacity to increase productivity through deep pessimism around its effect on jobs, to outright panic about its impact on society as we know it.
Uncertain horizons
In my new book Event Horizon Strategy, I liken the impact of AI to the uncertainty created by black holes in cosmology. Black holes are viewed as “singularities”: regions of infinite density and gravitational pull where nothing can escape. As objects approach a black hole, they reach a point called the event horizon. Beyond this, escaping the black hole’s gravitational pull becomes nearly impossible. What lies beyond the event horizon remains a mystery.
GenAI is currently proving just as mysterious but, just like a black hole, is simply too big to ignore. So how to adopt the right strategy? At best, getting it wrong risks missing out on opportunities and being left behind. At worse, it could mean the destruction of your whole industry or organisation. How can firms try and keep pace? How can they try and develop any kind of strategy that can harness this game-changing technology?
One solution is to look at the approach adopted by other firms. For Event Horizon Strategy, I reviewed the latest research on AI’s impact and conducted in-depth case-based analyses on OpenAI, Tesla, Palantir and other leaders in the field. I identify three strategies in the book and evaluate the pros and cons of each in addressing the singular threat and opportunity presented by the GenAI revolution.
Delaying tactics
The first strategy could be described as organisational inertia, where a firm makes minimal efforts to embrace the technology. I label this a “Circumnavigation and Delay Strategy”, when a firm decides not to directly engage with the technology but instead looks to continue with their existing models for as long as feasible. They circumnavigate the event horizon defensively and avoid being sucked into the black hole for as long as possible.
This strategy has been adopted by many traditional industries, such as those in the mining or construction sectors. Their business is not immediately at risk of being disrupted by GenAI, so they choose to simply maximise profits while they can.
Such firms can engage in a few AI projects that give some short-term advantages over rivals; however, they are unlikely to develop substantial organisation-wide capabilities or business units dedicated to AI. Instead, these companies are leveraging relationships with AI enablers, such as Palantir Technologies, or management consultants, such as McKinsey and Accenture. One risk here is whether these third-party providers are capturing most of the corporate value of GenAI through their high fees.
More significant is the fact that the traditional firms remain unprepared, and therefore ill-equipped, to deal with any seismic change that the evolution of GenAI might bring to their industry in the longer term.
Direct descent
At the polar opposite of this is what I term the Direct Descent strategy. As the name suggests this involves firms that have adopted a rapid and aggressive engagement with GenAI. Returning to the black hole analogy, such a swift descent is likely to result in destruction for most, with only a few surviving the journey.
OpenAI and Anthropic are two firms that have embraced this approach. Both are focusing on creating foundational Large Language Models (LLM). By rushing into GenAI, they are accelerating the industry towards an event horizon. They have benefited from an early mover advantage and the ability to play an outsized role in setting the standards in the sector.
Their business models focus primarily on charging consumers and firms for access to their LLMs. This approach differs from companies, such as Microsoft and Facebook, which might be developing foundational LLMs but look to make their money from using GenAI to enhance existing products, such as Office and Instagram.
The risks for OpenAI and Anthropic is that if they don’t diversify to develop products and services that leverage their LLMs, then they could end up in a race to the bottom, consumed by escalating capital expenditures and lower margins. Certainly, one or just a few companies will come to dominate LLM chat products. OpenAI and Anthropic may even come out on top. But the group of LLM providers are racing towards a future which results in little profit, unpredictable markets and will likely have only a few winners.
Flexible experimentation
There is also a third way: a strategy that offers a balance between survival and the capacity to exploit GenAI’s huge potential. This involves a strategic, phased investment in AI with a focus on learning and adaptability, with short-term gains that help build towards future developments.
Perhaps the most successful example of this strategy is Elon Musk’s Tesla Motors. While it might be best known for its electric cars, the company has been investing in developing “real-world AI” capabilities for over a decade. This AI strategy can be broken down into three main phases.
The first stage was the production of mass market electric vehicles with limited self-driving AI functionality. The importance of this phase was that multiple cameras on each vehicle were able to amass huge amounts of data. This data was then used to gain invaluable insights into real-world conditions and to conduct experiments in autonomy.
Tesla then used this learning in the second phase of its strategy: developing Full Self-Driving (FSD) capabilities. This culminated in the unveiling of the company’s fleet of robotaxis in October this year, with the promise of making human drivers redundant.
The third phase is the creation of “Optimus”. This is a humanoid robot that leverages the visual and movement systems developed in the FSD phase to perform a wide range of physical human tasks. Estimates put the potential market for this product to be at least tenfold that of the FSD vehicles.
The key to Tesla’s success has been the focus on an overarching direction or strategy, but also having the flexibility to adjust to the possibilities that arise from their exploration as they go. Each stage of the strategy is interconnected, with subsequent phases building upon the insights and technological advancements of preceding ones. This iterative process ensures the company has been able to remain flexible and forward-thinking in the dynamic AI landscape.
Next steps
For established companies, the key could be mobilising inhouse GenAI champions or superusers to accelerate adoption organisation-wide. In Event Horizon Strategy, I detail specific, GenAI use-cases that can be applied to workflows in HR, Finance and IT, before being scaled to innovative product development that can reshape business models. Applications in knowledge management systems, customer service automation and online experiments are low-hanging fruit that can help companies develop AI business capabilities before accelerating to the event horizon.
None of us can predict where GenAI is going to take us, but we do know we can’t ignore its potential impact. To cope with this uncertainty as well as exploit GenAI, firms might be best served adopting a strategy of deliberate experimentation.
By this I mean engaging in a sequence of strategic AI projects to identify high-value use cases that can then help improve productivity and profitability. The key isn’t just being flexible enough to exploit any short-term gains from these experiments, but to use learnings to prepare your firm for the uncertain future we all face.
“INSEAD, a contraction of “Institut Européen d’Administration des Affaires” is a non-profit graduate-only business school that maintains campuses in Europe, Asia, the Middle East, and North America.”
Please visit the firm link to site
