Management reboot essential for agentic AI strategy
pathdoc – stock.adobe.com
Agentic AI is not just an investment in technology with a fixed term replacement cycle, it can also be an employee that improves over time, says report
According to research from MIT Sloan Management Review and Boston Consulting Group (BCG), agentic artificial intelligence (AI)-based applications will lead to major management headaches. This is because technology purchases have traditionally been considered either as a substitute or a complement to human workers. Technology automates or augments and so can either be considered as a tool or as a worker.
The fact that agentic AI can act both as a tool and as a coworker breaks down traditional management logic, the authors of The emerging agentic enterprise: how leaders must navigate a new age of AI report warn. The report looks at how organisations now face an unprecedented challenge to manage a single system that demands both human resource approaches and asset management techniques.
For instance, the report points out that IT leaders look for predictable, scalable systems with clear technical specifications. Chief financial executives need investment models with measurable returns and depreciation schedules, while human resources executives require performance management frameworks and supervision protocols.
Sylvain Duranton, global leader of BCG X, the technology arm of the advisory firm, predicted that more money is likely to be spent on technology than on people. “When you look at the company of the future, it is pretty likely that the relative share of tech cost versus people cost will shift with a higher share allocated to tech costs.”
In the report, BCG and MIT Sloan Management Review warned that existing management principles are incompatible with how agentic AI is being deployed both as a tool and a worker. While tools scale predictably, workers adapt dynamically. The report’s authors note that agentic AI’s ability to do both simultaneously requires new organisational design principles.
Duranton noted that people management tends to involve a lot of considerations such as social engineering and negotiations with unions. “The same will come to the relationship with technology providers,” he predicted.
Another management change is that business heads need to assess the right time to invest in agentic systems and how these investments are made. According to BCG and MIT Sloan Management Review, business leaders are faced with balancing long-term capability building with short-term returns.
As the report’s authors point out, traditional tools require large upfront costs but deliver predictable returns through established depreciation schedules. Human workers, on the other hand, are an ongoing variable expense, but, as MIT Sloan Management Review note, their value appreciates with experience and training.
The report warns that agentic AI defies both models, requiring substantial initial development costs and ongoing variable costs, such as training models on new data. While many technology systems require ongoing maintenance, agentic AI systems simultaneously depreciate through model drift while appreciating through fine-tuning and emergent capabilities.
Duranton urged executives and IT leaders to rethink how they approach supplier relationship management. He said: “I think that it’s high time for many CIOs, and even the CEO and the C-suite, to strategise in terms of managing their portfolio of technology providers because these costs will be increasing over time.”
The fact that agentic AI evolves and develops over time means that value calculations fail since the most valuable applications have yet to be conceived. According to the report’s authors, conventional timing models applying conventional replacement schedules risk rapid value decay as systems fall behind the technological curve. This is because a traditional approach to upgrading the tech does not take into account for the speed of technological evolution.
Even if an AI system is able to deliver efficiencies right from the start, over time this could lead to a wider deployment, which has an impact on operational cost. Duranton recommended that IT and business executives assess the goal and intention of their technology strategy, as well as the portfolio of IT providers and products they use, to understand how much of the strategy relies on external technology providers, who may dictate the pace of innovation and future costs.
