A buyer’s guide to AI-powered legal research
AI can make legal research faster. The harder question is whether you can trust the answer.
Legal research is one of the most common uses of AI in the legal profession.
But a polished answer is not enough. It can still rely on the wrong authority, miss an important development or present something confidently that is simply incorrect.
So when assessing an AI-powered legal research tool, look carefully at what sits behind the answer, how easily you can check it and whether the tool helps lawyers apply legal judgement rather than bypass it.
In this guide, we will look at the questions worth asking when evaluating AI-powered legal research.
AI and legal research
Lawyers are now using AI across a wide range of legal tasks, from drafting and document review to summarisation and matter support.
69% of legal professionals use AI for legal research.
Legal research is the most common use case for AI, our recent survey found, with lawyers using AI to identify relevant authorities, explore points of law and get to useful source material more quickly.
But the risks are significant:
83% of legal professionals are concerned about inaccurate or fabricated AI-generated information.
In addition, 81% say they feel more comfortable using AI when it is grounded in legal sources.
That makes the quality of the underlying legal sources, and the ability to verify every answer against authoritative material, critical.
"No firm wants to be known for using AI badly. Once confidence is damaged, it can be very difficult to rebuild."
1. What is the AI actually researching?
The AI model is only part of the equation.
For legal research, the quality, authority and currency of the information behind it are just as important.
Ask the provider:
- What primary and secondary legal sources does the tool have access to?
- Which jurisdictions and practice areas are covered?
- Does it search the full text of those sources?
- How frequently is the underlying material updated?
- Does it distinguish between authoritative legal sources and general web content?
A strong model cannot compensate for weak or incomplete source material.
2. Can you see where the answer came from?
Legal research should be easy to verify. Lawyers need to be able to move from an AI-generated answer to the cases, legislation, commentary or guidance supporting it.
Ask whether citations link directly to the underlying source and whether the relevant material is easy to inspect.
If a lawyer has to start the research again simply to verify the answer, much of the time saving disappears.
3. Can you check the status and treatment of the authorities?
A citation is only useful if the authority can still be relied upon.
Ask how the platform helps lawyers check whether a case remains good law and understand how it has subsequently been treated.
That validation should sit naturally within the research process, rather than requiring lawyers to move between separate systems.
The same applies to legislation. Lawyers should be able to understand whether the provision they are relying on is current and relevant to the issue they are researching.
Access LexisNexis legal databases and real-time sources
Combine trusted LexisNexis legal content with real-time web sources to add wider context to your research, from legal developments to changing business, social and regulatory trends.
4. How does the tool reduce the risk of hallucinations?
No serious provider should pretend generative AI cannot make mistakes.
The better question is what has been done to reduce that risk.
Ask how the system is grounded, how outputs are tested and what happens when it cannot find enough reliable information to answer the question.
There is also a big difference between an answer that contains citations and one where the citations genuinely support the legal proposition being made.
Lawyers should be able to check that quickly.
5. Does it understand legal context?
Finding a case is not the same as understanding how important it is.
Good legal research involves working out which authorities matter, how they relate to one another and whether the law has moved on.
Ask how the tool handles:
- subsequent treatment of cases
- amendments to legislation
- jurisdiction
- conflicting authorities
- dates and recent legal developments
- commentary and practical guidance
The value of AI-powered research comes from combining speed with the legal context lawyers need to make a judgement.
6. How much work does verification take?
AI is usually sold on speed.
But the real test is how long the whole research process takes.
Give the tool questions your lawyers actually research and look beyond the first answer.
How quickly can they inspect the underlying authorities? Is it clear why a particular source has been used? Can they see the relevant passage? Can they move easily into deeper research?
The fastest AI-generated answer is not necessarily the fastest route to a reliable legal answer.
7. Does it fit into the wider legal workflow?
Legal research rarely ends with finding an authority.
The research may then feed into a piece of advice, a draft, a document review or further analysis.
Ask whether the platform allows lawyers to move naturally between those tasks while retaining the context of the original research.
Also test it with different users. A partner researching a difficult point of law may need something very different from a junior looking for a starting point. The tool should support both without encouraging either to accept the first answer uncritically.
What should you test during a trial?
Do not limit the trial to easy questions with obvious answers.
Use real research scenarios. Include recent legal developments, narrow points of law, questions with competing authorities and topics where the answer depends heavily on jurisdiction or date.
Then ask:
- Was the answer right?
- Were the right authorities used?
- Could the lawyer check them easily?
- Was their status clear?
- And did the tool genuinely save time once verification was included?
Those questions will tell you far more than a polished demo.
The bottom line
AI can remove a lot of the manual work from legal research.
But speed is only useful if the answer is built on sources lawyers can trust and check.
Before you buy, look beyond the model and the quality of the generated response and assess the legal intelligence underneath it.
How Irwin Mitchell uses Lexis+ with Protégé for legal research
Irwin Mitchell chose Lexis+ with Protégé for AI-powered legal research.
The firm wanted AI that could help lawyers get to relevant authority faster without losing the ability to check the underlying sources.
As Lynne Johnson, Practice Development Lawyer at Irwin Mitchell, puts it:
“Retrieval is brilliant with AI, but if you are retrieving information that you can't rely on and can't trust, then the AI becomes less useful overall.”
For Serious Injury Partner Deborah Rose, that means using Lexis+ with Protégé to explore a niche legal point, review the initial answer, open the cited sources and check the position before deciding what to do next.
Eleanor Windsor, Director of Knowledge Management, says:
“We chose to buy Lexis+ with Protégé because it's an AI tool that is grounded in trusted and managed source content.”
The aim is not to remove lawyers from the research process, but to get them to the right material faster and leave more time for judgement, strategy and client work.
Workflows that draft
Create outstanding drafts with workflows built on your documents and LexisNexis sources.
Pre-built & custom workflows
Move faster on everyday tasks and complex decisions with workflows that fit how you work.
Trusted content
Strengthen consistency with workflows that combine organisation knowledge and exclusive sources.
2026 research & thought leadership reports
In CTO we trust
Chief Technology Officers at leading law firms face a new challenge: not just deploying AI, but making it trustworthy at scale.
The Mentorship Gap
AI can help new lawyers to search, summarise, draft, and reason. But can it teach?
Integrating AI into legal workflows
Anything-goes AI use won't raise your standards. Legal teams need to restructure their workflows based on suitable AI.
Death of the rainmaker
Love them or loathe them, rainmakers are a law firm's bread and butter. But will AI-powered legal solutions see the practice take precedence over the partner?