AI and Change Management in Professional Services: From ‘Change and Forget’, to ‘Change, and Change, and Change…..’

The rise of AI in Professional Services is a formidable challenge for HR and change managers, targeting the very essence of human expertise, trust, and judgment that traditionally define these fields.

In many ways, AI is a continuation of a long history of technology-driven transformation, yet that process is one which professional services, with their deeply human focus, have been relatively untouched by – until now.

The argument outlined below is that AI is, in many ways, a different from a technology than those which firms are used to. This has a variety of implications, but here I focus on the change management which must accompany digital transformation.

Gen AI is a Better Knowledge Worker Than You

Unlike traditional technologies, it doesn’t simply address low skilled tasks through automation, but it can create, learn, and adapt, encroaching on higher-level human skills like creativity, decision-making and strategic thinking.

Three years into the GenAI revolution, AI is already disrupting knowledge-based jobs. Transcribing, as a job, no longer exists. It disappeared in 2023. Translation, as a job, has all but disappeared. These are jobs that are strongly associated with a specific task and thus disappeared quickly.

Most other professional services are more complex, and thus are being disrupted slower. However, AI now does many tasks better, faster and cheaper than many experts: qualitative analysis, creating dummy data, data extraction, content creation and so on.

At the most obvious level, this requires professional service firms to radically rethink the structure of work. However, deeper than this is a question of the professional identity of their experts: what does it mean to be a consultant, a doctor, a lawyer or an accountant?

Unlike previous tech, which focused on automating repetitive tasks, AI strikes at the heart of what senior professionals hold as uniquely human skills.

Higher Levels of Trust Are Likely to Be a Significant Enabler of Success

Second, AI is different because it requires greater trust between the firm and its workers than the usual top-down implementations. Unlike CRM or ERP systems, AI is inherently adaptable, evolving across different applications, which encourages (and, I would argue, sometimes necessitates) ground-up experimentation.

For example, my own research shows that over two-thirds of employees conceal their actual AI usage from management, highlighting a deep disconnect between the firms and employees. In the same vein, if, as is quite realistic, many of the AI experiments will result in fewer tasks, and potentially redundancies, then why should employees bother engaging in the first place?

Whilst some change managers might retort that ALL digital transformation projects require trust to succeed, that is not entirely the case. Of course, traditional tech implementations often require some form of co-operation to assist in requirements gathering, the tailoring of systems and so on.

However, most of these systems are all-but-finished products – they do not require the co-operation of individuals to actually experiment with different use-cases and builds. Of course, this is likely to change as the field matures. I’m personally approached every week by an excited entrepreneur who has built some ‘AI to disrupt the consulting industry’ and it’s reasonable to assume that not all of these will fail.

AI will become productized within existing and new software and SaaS, but crucially, many professional service firms have the internal data and expertise to actually develop some of these products better than anyone else. Deep expertise combined with an innovation strategy is already a winning strategy for some PSFs that have struggled in flat markets.

The Unprecedented Speed of Evolution

A third challenge for advocates of the human in technological implementations is the ever-evolving nature of GenAI. Compare what you have access to, free of charge, to what you had even two years ago. The speed of change is unlike anything outside of war.

In contrast, the internet took thirty years to mature, electricity took a quarter of a millennium. Of course, we are used to product updates, but can we really say that Windows 11 is a major change from Windows 10, or Hubspot in 2024 is significantly more capable than that three years earlier?

In contrast, AI has shifted from all but useless to miraculous in 3 years. In 2024, the frontier models have shifted from writing vaguely useful text to performing better than expert humans in fields ranging from cancer diagnostics to coding, as well as winning international competitions in art and literature. There is no reason to think this speed of change is slowing in the next 2-3 years. Change fatigue will become the norm if (it isn’t already) and recruitment will need to adapt.

This not only means that firms need to continuously manage change but they also need to be very aware of what is coming down the line. Being well informed about the potential of future AI models and their ecosystems should be feeding into strategic decision-making, from hiring decisions to decisions around productization and service development.

Implications for Firms and Experts

These challenges are not insignificant to any firm, but especially so in PSFs where identities are strongly tied to notions of expertise and experience. As data, AI, and automation take centre-stage, they’re shifting the balance of power in professional service firms.

The influence, once the stronghold of partners and senior experts, is now tilting towards those with analytical and data-driven roles, unsettling established hierarchies. With new power structures forming, firms must navigate the tension between preserving their expertise-centred model and embracing AI’s transformative potential.

The impact of AI is therefore existential: for many professionals, especially those whose roles centre on judgment, creativity, and decision-making, AI’s capacity to assist—or even replace—these functions is profoundly unsettling.

This isn’t merely about job roles being redefined; it raises fundamental questions about the purpose and value of human expertise in an AI-driven world: will the future knowledge professional be deskilled and there merely for oversight? If so, from where will they receive the skills if the lower grade jobs have been automated?

Anyone around in the Business Process Re-engineering years should have learned that successful change management and user involvement are critical, yet AI’s unique challenges will stretch traditional change management approaches to their limits. In short, the implementation of AI is not a ‘change and forget’ project.

In short, it is a ‘change and change’ project. It requires continuous exploration and exploitation – something that few firms outside tech have ever had to implement.