Kesha Tanabe has spent more than two decades in bankruptcy law, moving seamlessly between the public and private sectors. In the public sector, she has worked as an Assistant Attorney General, a Subchapter V Small Business Bankruptcy Trustee appointed by the U.S. Trustee Program within the Department of Justice, and she served as a U.S. Bankruptcy Judge for the District of Minnesota. Currently, Tanabe works at Vogel Law Firm, where she is a co-chair of the Firm's Technology Committee.  

In this conversation, Tanabe assesses the current state of technology in the legal field and how advancements in artificial intelligence are shaping the way attorneys work.  

You've tested a lot of AI products through your work on Vogel's tech committee. What's your overall take on AI in legal practice? 

Every law firm has some early adopters who can’t wait to try something shiny and new, and its fair share of lawyers who don’t even want a password or a free trial. But at our firm, we share a common commitment to providing great service to our clients. We try a lot of products, but it doesn’t matter how innovative a recent technology is if it does not directly translate to a better experience for our clients. 

To that end, I think we’re finding that AI is great for expediting certain tasks. It's great for automation. For example, I have been investing in and utilizing custom software applications that are tailored to my subject matter and practice area for over a decade. It’s great for analyzing data sets or executing form-driven tasks on a repeat basis. We also use AI in discovery when we have a large volume of information to process. This reduces “grunt work” and makes our work faster and more cost-effective. We can pass those benefits along to our clients. 

But AI can only do so much. In a legal context, it cannot “drive.” There is a big delta between what has been promised in pitches versus what these products can do in the context of real cases. I demo various products all the time, and I have yet to encounter a product that can replace the human element of lawyering. 

Why does speed matter so much in bankruptcy specifically? 

In our business, distressed companies are often referred to as a “melting ice cube.” If you cannot work quickly and efficiently, there will not be anything left to fight over. We need to help a company make payroll or manage a property on the verge of foreclosure. If we fail to develop a strategy and implement it quickly, value will be lost to both the debtor and its creditors. Bankruptcy lawyers must manage the very real risk that everyone will lose if we cannot work skillfully, quickly, and pragmatically. So, anything that might help us to be faster, more agile, or more cost-effective is an exciting possibility to a bankruptcy lawyer. 

How does the rise of AI compare to other technological shifts you've seen in your career? 

In my first year of law school, I had to physically go to the library and pull paper books off the shelf to complete my assignments. If someone did not put a specific volume of a treatise back on the shelf, you and your classmates might have to travel to a different law library and make a photocopy for the rest of the class. (Yes, that is a true story, but I digress.) 

By the time I got a Westlaw account, I thought it was going to change the world. And it did speed-up the process of legal research. But it did not render lawyers obsolete. Again, it eliminates some of the tedium of lawyering. Gone are the days of hourly contract attorneys sifting through document review for weeks or months at a time. But the human part of lawyering has not really changed much in my experience. 

What is that "human part" that AI can't replace? 

When you practice bankruptcy law, you are in constant conversation with other practitioners about the law and how it applies to real cases. A bankruptcy court is not a vending machine that spits out an order if you push a button. AI does not replace any of the negotiation, compromise, mediation, problem solving, or decision making in real cases– all of which are very human processes.  

One of the most basic obligations of a lawyer is to make sure that clients are giving informed consent. Clients need to know what is happening in their case and make informed decisions about how they want to proceed. This is a fundamental human dance between business owners or investors and their lawyers. A lawyer can have great technical skills and professional experience. But if a lawyer does not get to know his or her clients: their style, their approach to risk, what upside they're chasing, and what downside they're dreading, the lawyer really cannot effectively counsel a client when decisions need to be made. Large language models might be able to cobble together a memo, but they can’t replace the judgment of an experienced restructuring professional, and they can’t counsel each unique client who is trying to understand and evaluate their options in real time. 

Bankruptcy law seems to require a wide variety of skills. How are you able to adapt to the nuances of each case? 

I once read, "the world is a harmony of opposites." I've never forgotten it because it perfectly sums up the daily experience of being a bankruptcy lawyer. Sometimes a lawyer must be solitary for many hours at a time, doing focused work that is analytical, competitive, and task oriented. But this technical ability is not enough by itself. You can’t solve real problems without good listening, empathy, and interpersonal communication too. I admire lawyers who find a way to embody these seemingly opposite characteristics. 

How does that “balance of opposites” shape your view on the future of AI in law? 

I'm excited about making the practice of law faster and more cost-effective. But I am skeptical about whether AI will ever truly exercise judgment or counsel unique clients who are faced with challenges. So far, I just don't see AI changing this aspect of our practice yet. 

At the end of the day, the practice of law is a fundamentally human exercise. It's about communication, understanding people, and navigating complex situations. AI can help us do some of the grunt work more efficiently. But the advocacy, strategizing, and the ability to sit with someone in a vulnerable moment or a high stakes situation and help them to see a path forward—I doubt that can or will be automated anytime soon.  

Forming an LLC in California brings its own set of rules that shape how you operate. You might want the protection of a limited liability structure while keeping the day-to-day work manageable. You can make confident decisions when you understand the filings, naming rules, compliance habits and management responsibilities.  

Formation requirements and state filings  

You’ll need to file Articles of Organization with the Secretary of State. This filing creates your company as a separate legal entity and puts your chosen structure on record. After that step, you complete a Statement of Information within 90 days. This document lists your business address, your agent for service of process and your management structure.  

Many first time owners also draft an operating agreement early on. California does not demand that you file it, but keeping one in your records helps you clarify roles, ownership shares and financial expectations.   

For example, if you run the company with a friend, the agreement can outline who approves major spending or how you handle a partner’s exit.  

Naming rules and availability checks  

California requires your name to include ‘LLC’ or ‘Limited Liability Company’, and you cannot use words that mislead the public about what you do. You also need a name that no other registered entity in the state already uses. You can run your search through the state’s official BizFile Online portal.  

If your preferred name appears taken, you can adjust it by adding your city name, an industry term or a distinctive phrase. This saves you time and prevents rejected paperwork. Many owners also search domain availability at the same time so their web presence aligns with their legal name.  

Ongoing compliance and franchise tax obligations  

The state expects regular updates once you start an LLC in California. You file a Statement of Information every two years and pay the annual franchise tax. The minimum tax is $800, and if your revenue grows, an additional fee may apply.   

Setting reminders and checking your MyFTB account helps you stay ahead of these obligations, reduce penalties and maintain good standing.   

Staying compliant helps you avoid penalties that can cut into your operating budget. A predictable routine, such as reviewing your records each quarter, keeps your business healthy and reduces surprises.  

Legal considerations and management  

California treats an LLC as a flexible structure, but you still follow laws that protect your liability shield. You must separate personal and business funds, document major decisions and communicate clearly with any co-owners. If you treat the company as an extension of your personal finances, a court could challenge your limited liability protections.  

Strong management habits make day-to-day operations easier. For example, if you record decisions about new contracts or loans, you create a paper trail that supports you during disputes or audits.  

A cost-first view of everyday legal work: what gets cheaper, what stays expensive, and why the shift is more about budgeting than hype. 

For most people, “legal help” still comes with a single mental price tag: the hourly rate. In the U.S., that number frequently starts in the hundreds and rises quickly with geography, urgency, and perceived risk. Clio’s Legal Trends data—summarized publicly by Attorney at Work—puts the average lawyer hourly rate around $341 in 2024, with higher numbers in top-rate jurisdictions and practice areas.1 

That is not a critique of lawyers. The hourly model is a rational way to price judgment, accountability, and strategy. The problem is that consumers often meet this model at the exact moment they need something far more mundane: a clear structure, accurate language, and a document that does not accidentally create new problems. 

The mismatch has consequences. Some people overpay for routine drafting because they want certainty and someone accountable if things go wrong. Others do nothing, or copy a generic template from the internet and hope nobody challenges it. Either way, everyday legal work becomes an emergency purchase rather than a form of household maintenance. 

Consumer legal AI changes the economics less by “replacing lawyers” and more by lowering the cost of getting started. When drafting and revisions behave more like software than bespoke services, the first draft—and the tenth revision—stops feeling like a luxury. That shift alters who documents their life, how early they do it, and where expensive lawyer time is spent. 

Where the money goes in everyday legal life 

Ask any attorney what drives cost and you will hear a familiar list: interviews, issue-spotting, research, drafting, revisions, negotiation, filing, appearances. From the consumer side, however, the felt cost is often narrower and more frustrating: paying premium rates for what looks like templated writing. 

This is not about high-stakes litigation. It is about paperwork with consequences—tasks that are common, emotionally loaded, and easy to do badly. The law may be relatively settled, but the consumer’s writing can still be incomplete, inconsistent, or inadvertently damaging. 

Consider immigration support writing: character reference letters, hardship explanations, family or friend support letters, and “good moral character” narratives. These are not lawsuits. Yet they can feel existential to applicants and families, and the documents often need a specific tone, structure, and factual clarity. Consumers routinely pay hundreds of dollars for drafting help because a vague or disorganized letter can undermine an otherwise strong story. 

Then consider family and relationship paperwork: prenuptial and postnuptial agreements, temporary guardianship documents, co-parenting arrangements, and household agreements. Many couples delay them because the process feels expensive and adversarial. When they finally act, they pay a premium for speed and reassurance. When they do not act, they often pay later in a different currency: conflict, uncertainty, and reactive legal fees. 

Everyday contracts sit in the same zone. A room rental agreement between friends, a short-term vehicle rental arrangement, a basic loan or payment agreement, a bill of sale, an internship or employment offer letter for a small business, and a simple services agreement for a freelancer are all predictable sources of disputes when they are missing or poorly written. These are not “novel legal problems.” They are routine life transactions that become messy when they are undocumented. 

Finally, there is the first move in a conflict: the demand letter, the notice to cure, the written request for a refund, the letter of explanation that preserves facts before emotions harden into positions. These early documents are disproportionately valuable because they set tone and create a record. Yet they are exactly the kind of work consumers hesitate to pay for, even when a well-written letter prevents a small dispute from becoming an expensive one. 

How subscription Legal AI changes the price of drafting 

The most important shift consumer legal AI introduces is not better prose. It is a different pricing model for iteration. Traditional drafting has two prices: the unit cost of producing the document, and the ongoing cost of changes. Under an hourly model, consumers learn a simple lesson: every adjustment costs money. As a result, they revise too little. They send the letter without tightening facts. They sign the agreement without clarifying edge cases. They avoid the second question because the meter is running. 

A subscription model flips that psychology. Instead of paying again for each rewrite, the marginal cost of “one more document” and “one more revision” falls sharply. In unit-economics terms, the cost per document collapses, the marginal cost of the next document approaches zero, and the cost of iteration becomes negligible. That changes behavior: people iterate until the document is clear, because iteration no longer feels like a new invoice. 

The price points already exist in adjacent categories, which helps put the shift in context. Rocket Lawyer, for example, has been priced as a membership product; Forbes Advisor notes a Rocket Lawyer membership cost of $39.99 per month. 3 

LegalZoom’s Personal Attorney Plan lists pricing at $19.84 per month when billed annually (and $16.59 per month billed every six months), positioning ongoing access as “less than the cost of one hour with a traditional lawyer.”4 

Forbes Advisor describes prepaid legal services more broadly as a subscription-based way to access certain legal help at a fixed monthly or annual fee, often lower than paying a lawyer by the hour. 5 

Consumer legal AI extends the membership logic specifically to drafting. The practical result is a new cost structure: consumers can produce multiple routine documents in a month for the price of a predictable subscription, rather than paying a new drafting fee each time they need to correct language, add a clause, or rewrite a paragraph. 

This is also why the story is not “templates got cheaper.” Templates have been cheap for years. The differentiator is guided drafting: a workflow that prompts the user for facts, surfaces missing details, and generates a coherent document closer to what a lawyer would produce as a first draft. It reduces failure modes that come from consumers guessing what matters, and it makes iteration practical rather than painful. 

In practice, this category is often described as Legal AI: software that helps consumers generate and refine ordinary legal documents with guided structure and plain language, while still reserving true legal advice and representation for licensed counsel. 

Services like AI Lawyer—providing guided, AI-powered templates for immigration letters, family agreements, and everyday contracts—fit squarely into the “drafting plus structure” lane. The economic point is not hype; it is that drafting becomes a low-marginal-cost activity. 

For a concrete example of the guided-template model, see AI-generated legal templates for immigration, contracts and family law. 

“Template problems” vs “lawyer problems” 

One reason the cost conversation becomes confused is that “legal help” gets treated as a single category. It is not. There are template problems, where the law is relatively settled and the dominant failure modes are omission, ambiguity, tone, and factual disorganization. And there are lawyer problems, where the value is advocacy and strategy: assessing risk under uncertainty, negotiating against an adverse party, managing procedure, interpreting evidence, and taking responsibility for decisions that a consumer cannot safely make alone. 

Template problems include many immigration support letters, basic family agreements, everyday contracts, and first-step dispute communications. The stakes can still be meaningful, but the primary value is getting the structure and language right, making facts coherent, and avoiding self-inflicted confusion. 

Lawyer problems include criminal exposure, complex immigration matters, serious custody conflicts, high-value civil litigation, and regulatory investigations. Here, a clean document is not enough. The outcome is shaped by what is argued, how it is argued, and how risks are managed under time constraints. These are the areas where professional judgment and representation are not optional add-ons; they are the product. 

Consumer legal AI is economically honest when it stays inside the template-problem zone. Its value proposition should not be “you do not need a lawyer.” It is “you can arrive with a clean story, a coherent draft, and your facts organized.” That is not displacement; it is better inputs. 

For law firms, the efficiency implication is straightforward: if the first draft becomes cheap, the billable value of drafting alone compresses. The work that retains premium value is the work clients cannot safely do alone—issue-spotting, counseling, negotiation, and representation. In other words, the profession’s economic center of gravity shifts toward judgment. 

Why this is more than “it got cheaper” 

It is tempting to reduce the shift to “legal got cheaper.” The more interesting change is behavioral. When the cost of iteration collapses, consumers create documents earlier, document boundaries sooner, and preserve facts while memories are still fresh. They send the “here is what we agreed” follow-up after a conversation instead of trusting memory. They write the roommate agreement before resentment builds. They keep a timeline and supporting attachments instead of telling a vague story later. 

That alters downstream legal spending. People still hire lawyers, but they do so with better preparation: structured facts, a chronology, copies of messages, and draft language that can be reviewed and corrected. This reduces expensive time spent extracting information and increases time spent on analysis and strategy—the work that actually justifies premium rates. 

It also changes how legal risk is experienced. Under a purely hourly model, many consumers treat “making it official” as a luxury. Under a subscription drafting model, documentation becomes ordinary. More things are put in writing. More agreements get clarified. More narratives get refined. That can reduce the number of disputes that arise from ambiguity, not malice. 

Where savings are real—and where they are a dangerous illusion 

A cost-focused argument is incomplete without a warning label: cheaper drafting is not the same thing as lower risk. Consumer legal AI is a legitimate cost-saver when the primary task is drafting and organization. Standard letters, routine agreements, and early-stage dispute communications are natural candidates, especially when the alternative is no document at all. 

But there are scenarios where legal AI can create false confidence. If the matter involves criminal investigation, high-stakes immigration complexity, a serious custody fight, allegations of abuse, a large financial claim, or regulatory exposure, a well-written document is not enough. The consumer needs counsel who can evaluate risk, control communications, and provide representation. 

There is also a middle category: situations that look like template problems but are not. A prenup can be template-like in structure, yet become a lawyer problem with complex assets or cross-border issues. An immigration support letter may look simple, yet become a lawyer problem when it intersects with prior criminal history or allegations of misrepresentation. A demand letter can become a lawyer problem when it risks defamation, confidentiality breaches, or escalation into litigation. 

The safest line is simple: legal AI helps with preparation and drafting. A lawyer provides advice, advocacy, and risk-bearing representation. 

Access to justice as a budgeting problem 

Access to justice is often framed as a supply problem: too few lawyers for too many people. The consumer experience highlights another barrier: the entry price. For many households, the problem is not that lawyers do not exist—it is that the first hour is unaffordable, so the household never starts. 

The Legal Services Corporation’s Justice Gap research captures that barrier in stark numbers: low-income Americans did not receive any or enough legal help for 92% of their civil legal problems. Cost concerns are also a major driver of non-engagement—LSC reports that 46% of those who did not seek legal help cite cost as a reason, and 53% doubt they could find a lawyer they could afford.2 

Consumer legal AI does not “solve” the justice gap, and it does not replace the need for counsel. But it can reframe the consumer’s decision. Instead of asking “Can I afford a lawyer at all?” consumers can ask “Where should I spend real lawyer money, and where is a drafting subscription enough to get me moving?” 

In that framing, the economic story is not the end of the profession. It is a shift in the structure of spending. Drafting and iteration become cheaper, so scarce lawyer time can be reserved for the problems that require strategy, negotiation, and representation. 

What legal professionals should take from the shift 

For lawyers, in-house counsel, and legaltech teams, the cost shift has two practical implications. First, consumer willingness to pay premium rates for first drafts decreases as drafting tools improve. Value migrates toward review, customization under risk, negotiation, and representation—work where professional responsibility matters. 

Second, the client pipeline changes. As more people create baseline documentation early, lawyers may see fewer “I did nothing and now it is on fire” intakes, and more “I handled the basics and now I need help with the hard part” matters. That is a healthier starting point, even if it requires new habits in intake and document review. 

The most grounded way to describe consumer legal AI is also the least dramatic: it makes certain legal tasks behave economically like software. That does not end the need for lawyers; it clarifies where lawyers are irreplaceable. The headline is not that legal became cheap. The headline is that the cost of starting dropped—and in law, the ability to start early is often what determines whether someone pays a small amount now or a large amount later. 

Sources 

1. Attorney at Work – “Small Firm Hourly Rates by State and Practice Area” (summarizing Clio Legal Trends data, including a $341 average lawyer hourly rate in 2024).
2. Legal Services Corporation – The Justice Gap Report (2022): unmet civil legal needs and cost barriers.
3. Forbes Advisor – Rocket Lawyer review: membership pricing (noted as $39.99/month).
4. LegalZoom – Personal Attorney Plan pricing (listed monthly pricing and included benefits).
5. Forbes Advisor – Best Prepaid Legal Services: overview of prepaid legal subscription models.