Beyond the Term "AI Lawyer": Corporate Attorneys in the Age of Generative AI

Hello, I'm Noriaki Asato, Representative Attorney at LegalAgent.

What kind of support comes to mind when you hear terms such as "AI lawyer" or "generative AI lawyer"? In day-to-day corporate legal work, how to make use of generative AI is now being considered in contract review, legal consultations, due diligence in M&A, and other areas.

This development is a natural one. Generative AI is good at reading, summarizing and comparing text. Legal work, for its part, deals with large volumes of text such as contracts, minutes and statutes. In that sense, attorneys' work is a field well suited to generative AI.

However, some care is needed in how the term "AI lawyer" is understood. What is useful in corporate legal practice is not AI completing every legal judgment in place of an attorney. In my view, it is using AI for the work, and then having an attorney make the final, responsible judgment in light of the business background, risk tolerance, the state of negotiations, and the realities of internal decision-making.

I founded our firm as "a law firm for the age of generative AI." We develop our own AI Agent specialized in legal work, and our attorneys use it routinely in practice while providing legal services such as contract review, drafting, legal outsourcing, M&A, and startup legal work. This is an attempt to rethink the very way corporate legal services are delivered. In our daily work, we think separately about the tasks we leave to AI and the matters on which an attorney makes the judgment.

Misunderstandings Around the Expression "AI Lawyer"

When they hear "AI lawyer," some people may imagine AI answering legal questions, revising contracts and doing everything in place of an attorney. In corporate legal practice, however, that view has its dangers.

Taking contract review as an example, AI can explain what a clause means, point out issues of concern, and draft proposed revisions. For the work of identifying issues such as damages clauses, termination clauses and ownership of intellectual property rights, AI assistance is quite useful. However, it is the attorney who checks whether the output contains omissions or errors.

Corporate legal practice begins after those issues have been identified. You have to consider whether a given risk is realistic in light of the actual transaction, and whether it falls within an acceptable range given the transaction amount and profit level. Contract review only works in practice once you have taken into account circumstances such as how far to push for revisions given the balance of power with the counterparty, how much priority the business places on concluding this contract promptly, whether there has been trouble with the same counterparty in the past, and how well the internal decision-maker understands the risk.

AI cannot independently learn this context by looking only at the text of the contract. It cannot grasp information such as exchanges on Slack, the history of decisions in past matters, or the company's risk tolerance unless that information is somewhere it can refer to. That is why, when bringing AI into legal work, it is important for the attorney to fill in the background, add appropriate judgment, and return the result in a form the client can use.

Corporate attorneys in the age of generative AI need the ability to decide, building on AI's groundwork, what to raise as an issue, what to accept and where to negotiate. Speed in reading provisions alone cannot produce this judgment.

Working as an AI Native Law Firm

Ways of using generative AI include summarizing contracts, explaining clauses, and drafting reply emails and responses to legal questions. In every case, the basic premise is that the person in charge checks the content before using it.

There is, however, a clear difference between simply using generative AI and being an AI Native Law Firm. If you only use it partially as a convenient tool in the middle of your work, you have merely added AI to the conventional workflow. That approach can achieve some gains in efficiency, but the heavy burden of corporate legal work lies in the series of practical processes that go beyond one-off text generation.

Accurately understanding the request, gathering the necessary materials, and checking the differences between the latest and earlier versions of a contract. Grasping the intent behind the counterparty's comments and looking back at the history of similar matters with an understanding of the client's business and commercial flows. Referring to statutes and past decisions, and writing a risk explanation for internal use separately from reply comments for the counterparty. And finally, finishing the work to a standard that can be presented externally on the attorney's responsibility. Unless AI is built into this entire flow, its use in practice remains partial use of a tool.

The expression AI Native Law Firm that our firm uses indicates our policy of building AI into the work processes themselves; it does not mean any official legal qualification or certification. From the start of a matter through gathering materials, organizing the issues, and accumulating knowledge after completion, we aim to have AI and attorneys work while referring to the same background information. By creating such an environment, attorneys can secure time to engage deeply with the client's business strategy and negotiation challenges for growth.

The Limits of Tools in Practice and the Format of Deliverables

When introducing generative AI into legal work, the first thing many people try is pasting the text of a contract into an AI chat in a browser window. This is a natural starting point, and it is useful for rough summaries and checking issues.

In actual work, however, you hit a wall early on. In practice, contracts are normally exchanged as Word files. They arrive from the counterparty with tracked changes and comments; you, too, need to revise the provisions with minimal differences while keeping tracked changes, reply to the counterparty's comments, and prepare a separate explanation for internal use. Merely pointing out that "this clause carries risk" is not enough; the work is only complete once you have decided which words to delete, which phrases to add, and how to respond to the counterparty's proposal.

This is exactly why our firm developed Legal Agent, a Word AI Agent specialized in legal work. General-purpose AI is convenient too, but I feel that unless it connects directly with legal work processes, such as handling tracked changes and comments in Word and using different wording for internal and counterparty audiences, it is not easy to make full use of it in daily work.

In corporate legal work, the format of the deliverable itself has substantive value. What clients want is a draft contract that can be sent to the counterparty as is, or concrete options on which the company's officers can make decisions. From that perspective as well, thinking of it as "a corporate attorney who uses AI freely and sees things through all the way to negotiating with the counterparty and explaining internally" is closer to reality than the abstract label "AI lawyer."

The Domain of Attorney Judgment in the Age of Generative AI

Tools such as contract management and electronic contracting will no doubt continue to be useful in raising the productivity of the legal function. However, introducing tools alone does not automatically complete corporate legal work all the way to the final judgment. AI can assist with identifying points of caution, offering candidate wording, and finding similar clauses. But someone has to decide whether to adopt those suggestions, whether to present them to the counterparty, whether the company will accept a risk, and how to explain it to the officers.

In busy corporate legal departments, the burden that "a human has to check and fix each suggestion one by one" tends to remain. I sense that what companies without spare staff are looking for is not so much the features of a tool as legal capability that moves matters forward with a degree of responsibility while maintaining speed and quality.

That is why our firm emphasizes not only providing AI tools but also a structure in which our attorneys themselves use the AI Agent to provide legal outsourcing and AI-BPO. AI handles the groundwork, the attorney examines the legal risks and business impact and makes revisions, and we deliver work product that can be used in practice as is. This collaboration makes it possible to combine AI's processing speed with the attorney's responsible judgment.

I believe that, as generative AI advances, AI will take on much of the work of comparing contracts and preparing first drafts. Even so, I think attorneys will play the role of judging, based on AI's suggestions, where danger lies, which concerns are excessive, and what can be accepted as a business decision.

For example, in startup legal work, J-KISS, preferred shares and shareholders' agreements are interlinked. Even if AI can explain individual clauses, its advice falls short in practice unless it looks ahead to how that wording will affect future fundraising or sale negotiations in an M&A transaction. The same applies to M&A due diligence. Even if AI extracts change-of-control clauses and licensing issues from contracts and minutes, how to reflect them in the purchase price, the representations and warranties, and the post-acquisition integration process must be decided with the entire transaction in view. The attorney's job is to judge the impact on the transaction from the materials collected and reflect it in the contract terms and business response.

Points to Check When Choosing a Law Firm

When a company chooses an outside attorney or law firm, focusing only on whether they use AI is not enough. The first thing to check is whether the attorney in charge understands the realities of day-to-day corporate legal work. Unless the attorney can make judgments that take into account the type of contract, the background of the transaction, and even the interaction with the company's business divisions, AI output handed over as is will not stand up to use in the field.

The next thing to look at is which work processes AI has been built into. Is it just summarizing text in a browser, or is it linked all the way to tracked changes and comments in Word and drafting replies to the counterparty? Whether the firm has set up an environment in which AI can refer to past matter records, templates and internal rules greatly affects how practical the deliverables you receive are.

In addition, you cannot overlook who bears ultimate responsibility for the judgment. The key is whether an attorney with corporate legal experience checks the text output by AI and finishes it in a form that connects directly to business decisions.

"AI lawyer" is a friendly expression, but what our firm aims for is to build AI into our foundation and have attorneys well versed in corporate legal work deliver judgments that are faster and more closely aligned with the business. Rather than handing legal judgment over to AI wholesale, we aim to provide a legal function in which attorneys who make full use of AI support the growth of the business. The fees and scope of our arrangement in which attorneys use AI to take on day-to-day legal work are set out on the Legal Outsourcing page.