Legal research is more than just answers

AI has changed how quickly lawyers can do what they do everyday.

But courts, clients, and regulators are making something else clear: speed alone isn’t the standard anymore.

As Tej Kadiyala, AI Strategy Lead for Westlaw ANZ, put it during our recent webinar, Beyond answers: a new way to do legal research,

“AI has undeniably accelerated our output. But simply getting an AI‑generated response isn’t enough anymore. When the reasoning behind that response is hidden, the time you save generating it is often lost trying to verify it.”

This tension – between speed and confidence – is at the centre of legal research.

Output without traceability is the new risk

General‑purpose AI tools can produce polished, confident responses in seconds. The problem is what can happen next.

When reasoning is hidden, lawyers lose time trying to validate the output. Worse, subtle errors in jurisdiction, context, or interpretation can slip through undetected. For in‑house counsel, that can mean business decisions are made on unverified reasoning. In firms, it can become professional and reputational exposure.

Expectations may have changed, but defensible research hasn’t.

The fundamentals still matter: accuracy, currency, completeness.

The expectation now is that you can show the work:

• where the sources came from
• how the reasoning was applied
• what assumptions were made
• what the research does not cover

Courts are explicit about this. Transparency around AI use and accountability for outcomes are now table stakes.

From answers to accountability

James Jarvis, VP of Westlaw Product Management, described why this shift is happening,

“The way people have been working is a compromise. It is deficient and it is broken. That’s why we’re seeing reactions from courts and law firms – it goes back to the fundamentals of how legal work gets done.”

Legal research isn’t linear. It’s multi‑step, contextual, and nuanced. Tools that collapse that complexity into a single answer force lawyers to spend extra time reconstructing the reasoning.

One shift, two very different pressures

While the challenge is universal, the stakes differ depending on where you sit.

• In‑house teams need fast, accurate answers they can pass directly to the business – often across jurisdictions, often under time pressure.
• Law firms need research that can withstand scrutiny from partners, clients, courts, and regulators.

The takeaway

AI doesn’t lower the bar for legal work. It raises it.

As Tej summarised it:

“If you can’t trace it, you can’t trust it.”

That’s the new baseline.