Yes, the hype is real – AI agents really will change the face of work. Connected networks of singularly focused bots will create a step change in task automation and workflow efficiency in a way we haven’t seen since manual labour was revolutionised in the industrial age, and yet most large organisations will derive zero value from the truly incredible advances we are seeing almost daily.
The Competitive Response Window
Put bluntly, AI is available to everyone, will be used by everyone, and will therefore likely benefit no one. Let’s consider the digital revolution of the last 30 years. I vividly recall when sending an email had us staring at green text on a computer screen with our jaws on the floor. The same sense of wonder existed as we first started to search and find information on the internet, purchase goods online, receive personalised recommendations of what to watch on our TVs, and ask our cars to park themselves. But let me ask what you might think are a few silly questions. Who benefited more from email – Ford or Toyota? Who better monetised Google search – Nike or Adidas? Were data lakes more beneficial to Visa or Mastercard?
If Agents drive efficiency across industries, is it conceivable that one energy company might choose to deploy them while an equally sophisticated competitor might prefer to stick with higher cost manual processes?
Agents will become a matter of course. In the same way that my kids simply can’t imagine a life before smart phones, the Agentic future will soon be normalised, and the associated efficiencies will be taken for granted. How therefore does an incumbent in a competitive market create advantage?
The answer perhaps lies in the ability to act with speed. While it is unlikely that one bank will rapidly deploy Agents and win market share from all the other banks so that they are the last bank standing, speed might nevertheless create short lived competitive advantages that shift market share at the margins. There is a competitive response window that incumbent organisations are eyeing with nervousness, but let’s be clear… it is unlikely that American Airlines will triumph over or be subsumed by British Airways by virtue of faster deployment of Agents.
The Curse of Size
For most industries, incumbents acting at speed is an oxymoron, and the urgency around AI is nowhere near as keen as it might otherwise be as a result of knowing that my fellow incumbents are equally incapable of acting at speed as we are. So, we convene a few AI taskforces, read a bunch of articles, perhaps even go to a conference or two. We create a pilot here and there, hire in an external organisation to wow us with some demos, and wait till Agents have been so thoroughly deployed across hundreds of organisations that the inevitability of our Agentic future requires zero risk.
No one looks back at the internet revolution and remembers the companies that banned internet usage by their employees. We might laugh at the thought, and it really did happen quite widely, but in time these things self-correct. Most incumbents will feign urgency around AI but ultimately wait for the currents of change to wash over them and for autonomous Agents to simply become part of the furniture.
No urgency until it is too late
As long as Agents drive efficiency or more effective ways of completing known tasks, incumbents will breathe easy in knowing that forward progress towards monetising AI is hard to sustainably compartmentalise within one organisation and not have it applied in another. However, a colleague introduced me to a quote from a French innovator called Marc Giget who said “No candle-maker has become a bulb manufacturer, no carriage-maker has become a car producer, and the post office did not invent the email”. These examples were more than evolution or mere efficiency gains but rather, a complete step change in the very nature of an industry. This is perhaps where things become a bit more urgent.
Will Agents represent more than the automation of individual tasks? Could these efficiencies fundamentally shift customer value propositions, service models or the nature of products we take for granted? If Agents create new possibilities that change the foundations of how industries operate, maybe the race is a little more urgent than large organisations realise. For example, what if AI automates the task of grocery shopping, gardening, going to the doctor for an annual health check or needing an estate agent to sell your house? If Agents can disrupt entire industries rather than just automate tasks, then we are no longer in a world of zero sum gain across competitors. There will be distinct winners and many losers.
Some industries are rightly worried about the Agentic future. AI could well represent the destruction of highly profitable business models. When Blockbuster died, the underlying customer need didn’t change, but the way of fulfilling that need transformed. Anyone in a services industry should take heed. Consultants, accountants, financial planners and lawyers are in the crosshairs. Like all cautionary tales of disruption, there is no urgency until it is too late.
Running the wrong race
Irrespective of whether Agents will drive efficiencies or radically re-write industry norms, many incumbent organisations are exploring AI in a predictable but unhelpful way. They are often focused on the tech, or on the policies needed to limit misuse. We have seen over the last 20 years that the obstacles to monetising digital and data have very little to do with technology, but on the question of policies, AI has undoubtedly more potential for harm then any technological advance we have seen in recent times. Should organisations therefore be spending lots of time and energy trying to limit the misuse of AI? How effective were value statements or policies when it came to the unauthorised accounts and credit card applications at Wells Fargo, the false emission readings at Volkswagen or the government tax scandal at PWC?
I am not suggesting that Agents should be developed with little thought to ethical considerations. However, policy in of itself will not curb misuse. Rather, we should be focusing on our overall readiness for the Agentic future. This will require a significant change to corporate culture, where control and hierarchy provide the scaffolding for how we organise. Agents will challenge traditional approaches to organisational structures and culture more than anything we have ever experienced. Even during the industrial revolution, the people being organised to do the work in completely different ways were still people.
The ultimate blocker to effective deployment of Agents will be operating models and culture, not technology and policies. The latter two obstacles might be the excuse used to slow down the embracing of Agents, but it will ironically be the people who create, interpret and interact with Agents who will need the most attention. Focusing predominantly on use cases and the anticipation of new value pools is akin to running the wrong race.
***
Most industries will see mouth-watering opportunities across their value chains and Agents will represent never before seen efficiency pools. Incumbents in these industries will all benefit, and the rising tide will make it hard for any one organisation to win at the expense of another. However, some industries are more susceptible to revolution – think of the massive shifts in business models that have impacted newspapers, taxis and movie cinemas. Incumbents in these at-risk industries can’t afford to be as complacent when it comes to their Agentic future. Irrespective of whether facing an existential threat or the mere re-writing of industry economics, speedy progression towards the Agentic future requires a focus on the very challenges that have been most difficult to overcome when it comes to digital and data. Extracting benefit from Agents means embracing a shift to operating models and culture. The irony should not be lost that the only way to successfully replace people with Agents will be to focus on people.