Amara’s Law states that we tend to overestimate the effect of technology in the short run and underestimate it in the long run.
Right now, we are at an emergence point between these two states, and we are in danger of missing the bigger picture. Businesses are underestimating what they already have; and as AI develops further, they are underestimating how they will be able to use it.
As we enter a world where AI can provide super intelligence on any topic, and where the cost of that intelligence continues to decrease, it’s time for business leaders to ask themselves how they are going to leverage the immense power of AI to drive their organisation forward.
To remind ourselves of the speed of change we are experiencing, it is worth noting that the first public version of OpenAI’s ChatGPT model was released less than three years ago. Since, foundational models have spurred an exponential growth in our capacity to analyse data and bring value like never before. The development of small language models is now enabling businesses to create bespoke solutions that are precise, accurate and controllable – and cost much less than foundational models, which require more resources.
In this next stage of adoption, AI is adapting to connect with people better. We are already seeing the rise of provincial AI models in, for example, Saudi Arabia, France and Japan that are specific to their language and culture, while embodied AI puts AI into machines, turning them from chatbots into something real that can interact with us in the physical world. Think the Internet of Things, but exponentially more powerful.
Take AI agents. This is not just about process automation or streamlining existing workflows; this is AI thinking about a solution, planning that solution and acting on that solution, maybe without humans even being involved.
Autonomous AI agents can also increasingly understand physics and the physical interactions of people, machines and the world around them, and the next wave will empower even bigger thinking.
We are already seeing the benefits of this within PwC. We built our agentic AI platform agent OS out of necessity because we were creating so many agents to serve our client work, but they were siloed and difficult to supervise and maintain at scale. Now we see agent OS as a fundamental tool in everything we are building, both internally and in the service of clients. Agent OS is enabling my team to work in five-day idea-to-production ready sprints, delivering more output at a higher quality in a regulated environment and achieving more than I ever thought possible.
That kind of timescale is staggering. And crucially, not at the expense of our workforce. We are not getting rid of engineers, we are hiring more, because we can produce more at higher quality and faster pace, enabling us to support our clients in a fundamentally different way.
For businesses there is an urgency to understand and embrace all of this. We are rapidly moving past predictive and prescriptive AI to an autonomous environment - and the winners will be those who are successfully able to infuse this throughout their organisation.
That means focusing on autonomy, speed and scale. This is not about automating ten steps of a process; it is about getting an agent to autonomously decide how to achieve the output you ask for. Scale is also important. If you are building agents separately in silos, you are building them wrong because agents can be general purpose and used everywhere. Speed matters too. If your business is still thinking in five-year roadmaps, you need to rethink that.
If all that sounds daunting, the good news is that AI can build AI. AI is very good at software development, and it will become even better as it uses the physics of the world to generate 3D models, CAD ( Computer-Aided Design) designs, design medicine and create new materials.
For business leaders, therefore, the task is clear. They will need to be cross-disciplinary, as the need increases for people who understand both the business and AI, enabling them to lean in to leverage AI and empower them as leaders.
Leaders must also build trust into the design of their AI systems and put guardrails in place to ensure that AI is being used in an ethical and responsible way. AI can’t be adopted if it’s not trusted by those using it.
They must also take their workforce with them on this journey and understand that the education and training they provide must be different from what has gone before. This is not about providing a few courses for people to do on their computer, this is hands-on learning and understanding through doing.
Businesses need to change their expectations of what is possible – because the expectations in the market are also changing quickly. Not just what their organisations could be doing today, but what they can and should be doing tomorrow with the tools of this new world.