As regulatory complexity accelerates, Regulatory Affairs leaders face a hard question: can existing teams keep pace through more effort alone, or does the function now need to be redesigned around intelligent, governed automation?

For many Regulatory Affairs teams, the pressure point is no longer knowledge. It is capacity. Global requirements are expanding, guidance is shifting across jurisdictions, and teams are spending days each week simply identifying what has changed before they can assess what it means.

Caroline Shleifer, Patricia Teysseyre and Anam Mukhtar of RegASK argue that this is where the next phase of artificial intelligence becomes significant. Generative AI has already shown value in summarising documents, drafting content and accelerating regulatory intelligence. But the deeper opportunity lies in agentic AI: systems that can move beyond responding to prompts and begin orchestrating regulated workflows within defined human-approved boundaries.

Generative AI has become a useful productivity layer for regulatory teams. It can condense lengthy guidance, compare requirements across markets, assist with drafting standard operating procedures, and support document review. In an environment where regulatory professionals are managing high volumes of information across multiple jurisdictions, these capabilities are valuable.

But they do not, on their own, solve the core operational challenge.

Regulatory work is rarely a single task. A change detected in one market may trigger impact assessment, labelling review, stakeholder notification, dossier updates, quality documentation and audit trail requirements. A tool that summarises information is helpful. A system that can connect the steps, initiate the right actions, route outputs for review and preserve evidence of decision-making is something more strategic.

That is the promise of agentic AI.

Rather than acting as a passive assistant, agentic AI can operate as an active workflow partner. It can monitor, analyse, draft, escalate and record activity across systems, while qualified regulatory, quality and pharmacovigilance professionals retain decision-making authority.

Why the Regulatory Affairs Function Is Under Pressure

The case for change begins with the scale of the workload. Regulatory teams are expected to monitor evolving requirements, interpret local implications, coordinate internal responses and maintain inspection-ready records. This must often occur across markets that differ in language, regulatory maturity, enforcement posture and submission expectations.

The traditional response has been to add people, increase manual checks or rely on fragmented digital tools. But that approach has limits. Complexity is scaling faster than headcount.

The article points to a growing risk for organisations: regulatory intelligence can become a bottleneck before any scientific, clinical or commercial decision is made. If teams are spending a substantial share of each week on monitoring alone, less time remains for strategic interpretation, cross-functional advice and early risk management.

For life sciences organisations, this matters because delays in regulatory interpretation can affect market access, labelling, compliance readiness and product lifecycle decisions.

What Agentic AI Changes

Agentic AI can be understood as a governed system of digital agents designed to work towards defined goals. These agents can perceive changes, reason across data sources and trigger actions within agreed limits.

In practical terms, this could include:

• scanning health authority websites for relevant updates

• identifying whether a change affects specific products or markets

• preparing a first-pass impact assessment

• drafting a memo or updated procedure

• alerting regional teams

• routing content for review and approval

• preserving source provenance and audit trails

This does not remove professional judgement. It changes where that judgement is applied.

The regulatory professional moves from manually searching, copying, comparing and formatting to supervising, validating and deciding. The value shifts from task execution to risk interpretation, governance and strategic advice.

Early Use Cases Show the Potential

The article outlines pilot-style examples that show where agentic AI may deliver measurable gains.

One example involves global labelling change management. A company monitoring changes across 25 markets used an agentic workflow to detect labelling-related updates, assess relevance to the product portfolio and prepare draft impact memos for human review. The reported benefits included faster impact assessment, fewer missed updates and reduced version-control errors.

Another example focuses on Module 1 dossier assembly. In this case, the system populated country-specific templates from internal systems, assembled required documents and prepared packages for electronic submission, subject to validation and human approval. The result was a reduction in cycle time and fewer administrative rework errors.

The broader lesson is not that AI should take over regulatory work. It is that repetitive, rules-based, high-volume workflows are increasingly suitable for AI-enabled orchestration, provided governance is built in from the start.

Governance Is the Deciding Factor

The most important question is not whether agentic AI can act. It is whether organisations can control how it acts.

In regulated environments, any AI-assisted workflow that touches submissions, labelling, standard operating procedures, quality systems or pharmacovigilance must be designed around accountability. Human oversight, validation, documentation, access control, version management and explainability are not optional.

The article highlights several governance principles that should shape implementation:

• AI-generated outputs should support decision-making, not replace accountable human judgement.

• Regulatory and scientific outputs must be reviewed and approved before being relied on or submitted.

• Predictive insights should be treated as decision support, not as definitive forecasts.

• The system must preserve source provenance, audit trails and change history.

• Controls should be proportionate to the risk and context of use.

This is where agentic AI becomes a leadership issue as much as a technology issue. Poorly governed automation can increase risk. Well-governed automation can improve consistency, transparency and audit readiness.

Start Small, But Design for Scale

The article makes a practical point: Regulatory Affairs leaders do not need to wait for enterprise-wide transformation. A more realistic approach is to begin with one high-volume workflow where the current burden is clear and the risk controls can be defined.

Good starting points may include:

• horizon scanning

• labelling change monitoring

• dossier assembly

• SOP gap assessment

• multilingual regulatory intelligence review

A focused pilot should include clear metrics, such as time to action, missed update rate, review burden, rework, audit trail completeness and hours redirected to higher-value work.

The strongest pilots will not simply test whether the technology works. They will test whether the operating model works.

Practical Takeaways for Leaders

1. Treat AI as an operating model change, not a tool upgrade.The real value lies in redesigning workflows, decision points and accountability structures.

2. Prioritise high-volume, rules-based processes.Agentic AI is best introduced where work is repetitive, measurable and currently resource intensive.

3. Keep professional judgement at the centre.AI can prepare, analyse and escalate. Regulatory, quality and pharmacovigilance professionals must remain accountable for decisions.

4. Build governance before scale.Validation, audit trails, access controls, source traceability and documented review pathways should be designed into the pilot, not added later.

5. Measure strategic value, not just efficiency.The goal is not only faster drafting. It is improved compliance readiness, better risk visibility and more time for expert judgement.

The Strategic Implication

Agentic AI will not remove the need for regulatory expertise. It is more likely to increase the value of that expertise by reducing the manual burden that currently sits around it.

For Regulatory Affairs leaders, the opportunity is to move from reactive compliance management to proactive regulatory execution. That shift will require careful governance, cross-functional confidence and a willingness to rethink long-established workflows.

The organisations that succeed will not be those that automate the most tasks the fastest. They will be those that use AI to strengthen accountability, improve oversight and give skilled professionals more time to focus on the decisions that matter.