Consultants Are Already Using AI. Almost No One Agrees What Consulting Is.

Before consulting can govern AI use, it may first need to define itself.

Lately, I’ve spent a great deal of time debating a deceptively simple question: What exactly is a pharma or biotech consultant?

The more conversations I have, the broader the definition becomes. Consulting can start right out of school or evolve from decades of industry experience. It exists inside large firms, boutique networks, and solo practices alike. Some consultants provide temporary operational support; others serve as strategic advisors, fractional executives, or board members. Even Big Pharma and biotech consulting, while closely related, often operate under very different assumptions, timelines, and delivery models.

A recent set of conversations reinforced this for me. One was with Dean Molloy, CEO of NPG, a consulting organization with hundreds of consultants operating at institutional scale. Another was with Riley Elmer, President of USC’s Advanced Degree Consulting Club, representing a new generation entering consulting through the traditional pharma career pipeline. My own work sits primarily in the biotech ecosystem, which often operates with different resource constraints, risk tolerances, and operational realities. Yet all three perspectives exist within the same broader industry conversation while reflecting very different visions of what consulting can be.

It made me realize that consulting is not defined along a single axis. At minimum, the profession spans four overlapping dimensions:

• Expertise — What are you great at?
• Authority — What level of responsibility do you carry?
• Delivery Model — Are you a sole practitioner, boutique firm, or scalable institution?
• Customer Context — Who are you serving?

That ambiguity mattered less when consulting was primarily relationship-driven and operationally opaque. AI changes that. Suddenly, clients are beginning to ask harder questions about how work is produced, reviewed, and validated. Consultants, clients, and vendors are all using AI in some form—drafting documents, summarizing data, generating analyses, and accelerating workflows. Much of this use is informal. Few organizations have formal policies. Fewer still have defined what acceptable use actually means.

This creates a fundamental tension. If the profession itself is loosely defined, how do we define acceptable AI use within it? Is a consultant’s value in the output, the process, or the judgment behind it? If a deliverable is partially generated or informed by AI, what does that mean for ownership, accountability, confidentiality, and quality?

These are not theoretical questions, particularly in regulated industries. Consulting work often relies on confidential client information, proprietary operational knowledge, and experience-driven judgment. AI introduces real risks: hallucinated outputs, inappropriate data exposure, and overreliance on systems that were not designed for regulated decision-making. At the same time, ignoring AI is not a viable strategy. The efficiency gains are real, and the competitive pressure to adopt is accelerating.

AI is not creating a governance problem in consulting—it is exposing one that already existed. The profession has operated for decades without a shared definition, standard set of practices, or broadly accepted code of conduct. Before consulting can establish meaningful standards around AI, it may first need to define what consulting actually is.

That is the purpose of this series. Not to advocate for or against AI, but to explore where it fits, where it creates unacceptable risk, and what professional accountability should look like in an AI-enabled consulting environment. Because if consultants do not define these boundaries themselves, clients, regulators, and contracts eventually will.

Questions to Consider

• Should acceptable AI use differ based on a consultant’s field of expertise?
  For example, is AI-assisted drafting in marketing strategy equivalent to its use in regulatory CMC, quality investigations, or clinical decision support?
• Does the level of authority and responsibility carried by a consultant change what constitutes acceptable AI use?
• If AI contributes to a deliverable, where does accountability ultimately reside: the consultant, the firm, the client, or the system itself?

For Discussion

The discussion around AI in consulting is often framed as a technology debate, but it may actually be a governance and professional identity problem. Consultants operate across different domains of expertise, levels of authority, delivery structures, and customer environments—yet the industry often speaks about “consulting” as though it were a single, uniform activity.

AI intersects with each of these dimensions differently. A temporary contractor using AI to summarize meeting notes is not operating under the same expectations as a fractional executive advising a board.  How should we approach each use case?  

The goal of this series is not to advocate for or against AI, but to begin a more disciplined discussion around where AI fits within professional consulting, where boundaries may need to exist, and how accountability should evolve alongside these tools.