June 2026

AI is here. The question is whether you’re using it right?

AI has arrived in legal practice. The bigger question now is whether legal teams are using it in the right places and with the right level of oversight.

In our recent CPD webinar, David Curtin spoke with Rok Popov-Levinsky, founder and CEO of NPL Legal Tech Advisors, about what is happening inside legal teams as AI adoption accelerates and where the risks are starting to emerge.

Some key themes stood out.

Legal teams are starting in the wrong place

Many teams are trying to apply AI to the hardest legal work first. For example, complex reasoning, strategic advice, and nuanced judgment calls. That’s where AI is currently least reliable.

Where AI is proving genuinely useful is in structured, repeatable work, like intake triage, first-pass contract review, and standard drafting. The kind of operational tasks that can be clearly documented, broken into steps, and repeated consistently across a process.

AI hype is creating confusion

Agents are the latest major buzzword in legal tech, with vendors rapidly promoting systems that promise to automate increasingly complex workflows.

The problem is that not all agents work the same way, and not all carry the same level of risk. Some assist with structured workflows where a human still designs each step. Others operate autonomously, making their own decisions across systems with very little human review in between. That is a significant difference and one that’s getting lost in the noise.

AI isn’t replacing lawyers at scale

What the labour market data shows is cost-cutting dressed up as AI-driven innovation: firms announcing redundancies in the name of AI efficiency when the technology is not yet capable of delivering those gains.

That narrative is not helping anyone. It creates fear in the teams trying to adopt AI and sets unrealistic expectations for those selling it, while the real opportunities get overlooked.

AI delivering breadth and speed

In a recent pro bono case, a legal team used AI to review all 100 briefs they received. Something that simply would not have happened before.

That’s what AI augmentation looks like in practice. Not replacing the lawyer, rather allowing them to do the job more completely than was ever previously possible.

The standard is also shifting in how in-house teams evaluate external firms. What was previously an acceptable four-day turnaround on a draft is becoming unacceptable in some markets because competitors can now deliver the same quality in a day.

Human checkpoints still matter

AI works best as an augmentation tool, not as a replacement for human judgment.

One legal team built an autonomous agent with access to their internal systems. It hallucinated, deleted files, and there was no backup.

The lesson is simple. The more autonomous the workflow becomes, the more important human checkpoints become. Particularly before anything external, irreversible or high risk occurs.

The teams making the most progress are experimenting safely

The legal teams making the most progress with AI are not necessarily the ones with the biggest budgets or the most advanced systems. They are the teams willing to experiment carefully, start small, and treat failed tests as learning opportunities rather than setbacks.

For many in-house teams, that means starting with low-risk administrative or process-heavy work using tools already available within the organisation. Get comfortable there first. Then build from it.

The full webinar examines these issues in greater depth, offering practical guidance on AI adoption, risk management, and where legal teams can most effectively begin.

Watch the full webinar below: