The Most Important and Overlooked Estate Planning Questions That AI Does Not Properly Address
ParkBridge Wealth Management Fall Webinar Series | September 24, 2026
Host: Jonathan I. Shenkman, President & Chief Investment Officer, ParkBridge Wealth Management (jonathan@parkbridgewealth.com)
Speaker:Avi Z. Kestenbaum, Esq., Partner, Meltzer, Lippe, Goldstein & Breitstone, LLP (akestenbaum@meltzerlippe.com)
Jonathan Shenkman: Good morning, happy holidays to those who are celebrating this month, and welcome to the ParkBridge Wealth Management Fall Webinar Series. This program is entitled "The Most Important and Overlooked Estate Planning Questions That AI Does Not Properly Address." As always, my name is Jonathan Shenkman, and I'm the President and Chief Investment Officer of ParkBridge Wealth Management.
In that role, I serve in a fiduciary capacity to help my clients achieve their financial objectives. The goal of my programs is to bring professionals together to help them better serve their clients. I do this by educating attendees on the latest topics in wealth planning and by encouraging collaboration among a client's attorney, CPA, and financial advisor where appropriate.
I focus on working with high-net-worth families, businesses, and not-for-profits. I manage individual investment portfolios, trust accounts, corporate retirement plans, and endowments. In addition to the 20 or so events I run every year, I write about investing and financial planning for a variety of periodicals, including Barron's, CNBC, Forbes, Kiplinger, The Wall Street Journal, and Trusts & Estates magazine.
You can see all my work at ParkBridgeWealth.com/articles or by following me on social media @JonathanOnMoney. You can also check out my weekly podcast, also called Jonathan on Money, on Apple, Spotify, or wherever you get your podcasts. Finally, my first book, D Is for Diversification: The ABCs of Personal Finance, is available on Amazon or at JonathanOnMoney.com. It's a great way to support these programs.
Today, we're privileged to hear from Avi Kestenbaum of Meltzer, Lippe, Goldstein & Breitstone, based on Long Island, New York. Avi is co-chair of Meltzer Lippe's nationally ranked trusts and estates practice and chair of the firm's tax-exempt organizations practice.
He advises CEOs, ultra-high-net-worth families, multinational businesses, and major nonprofits on sophisticated domestic and international tax, estate, and asset preservation matters. Avi is widely recognized for his big-picture planning philosophy, his work in complex family business succession, and his success representing clients in IRS and state tax audits.
He has developed a niche national practice in charitable planning and nonprofit structuring and is co-founder of STEP Long Island. Avi is a prominent national lecturer and prolific author whose work appears in leading tax and estate planning publications, and he's frequently quoted in major media outlets. He serves on the editorial board of Trusts & Estates magazine, is an ACTEC Fellow, and has taught tax law at Hofstra University.
With that introduction, I'll turn the program over to Avi.
Knowledge Versus Wisdom
Avi Kestenbaum: Thank you very much, Jonathan, as always, and thank you for providing so many wonderful educational programs to all of us.
Let's start with a concept: knowledge versus wisdom. If you were getting advice about your marriage, your children, or your emotional life, would you want to go to the most knowledgeable person or the wisest person? Put differently, which matters more to you: knowledge or wisdom? I think most of you would say wisdom.
Here's a dictionary definition. The fundamental difference is that knowledge is the accumulation of facts, information, and skills, while wisdom is the ability to discern, judge, and practically apply that knowledge to make sound life decisions. In short, knowledge is an asset of the mind, whereas wisdom is a virtue of judgment.
Knowledge is about data, the facts gained through study or experience. Wisdom is a deep understanding and the judgment to apply those facts effectively. Wisdom comes from trial and error, life experience, self-reflection, and time. It is also more practical than knowledge, because it answers the why, when, and where. Wisdom includes judgment, discernment, foresight, and prudence. I could go on, but you probably understood this intuitively before I started.
As always, I'm now down to 25 minutes, so I'll speak quickly and give you the broad strokes and the most important points. There can always be follow-up.
Let me ask you another question. What makes one doctor, lawyer, or accountant better than another? Take a group of very experienced professionals, with 10, 20, or 30 years in their field. Let's say they have a similar knowledge base. Why choose one over the other?
You might say you like one's personality better, or that one is more empathetic. But most important is judgment. We all have the knowledge. What matters is how to apply it, and with what creativity. I'd put creativity under wisdom and judgment too. It goes beyond knowledge.
Where AI Falls Short
Now we can start to understand the weakness of AI, specifically in estate planning. I'm not here to judge AI in other areas. AI can be amazing, and it could also be horrific. People today aren't sure whether to be thrilled or to worry that we'll soon be living in a Terminator movie. That's how broad the spectrum is.
I'm not here to comment on any of that, and I don't know anything about it. As my wife and children will tell you, when it comes to technology, including AI, I'm the last person to ask.
In fairness to AI, though, many of us just finished the Day of Atonement, and I didn't want to embarrass it. AI may have human-like features at this point, so I asked it directly. I went into one of the more prominent AIs (I won't say which one) and typed: "Does AI lack wisdom?" Why embarrass it without giving it a chance to respond?
This was its response: "Yes, artificial intelligence completely lacks wisdom. While AI excels at gathering, processing, and generating massive amounts of knowledge, it cannot exercise true discernment, empathy, or moral judgment."
It then listed its deficiencies, which we'll apply to estate planning shortly:
- AI has no lived experience or consciousness. It has never lived and never experienced anything.
- It lacks context and common sense.
- It cannot navigate moral and ethical dilemmas.
- It mimics, but it does not discern. It simply retrieves other people's words and information.
AI doesn't feel, and it doesn't understand the concepts. It's literally spitting back information.
Now, I don't want to say anything overly controversial, and I don't mean this negatively. I have members of my own family who may be on the autism spectrum. But if you went to a really bright person with autism or Asperger's, that person could be a wealth of information. Because there may be a lack of understanding, of empathy, of the emotions behind it, you probably wouldn't go to that person for advice. I don't mean to be critical, and I believe we all have a touch of that in us. I think we're all somewhere on that spectrum. But the point stands, and I'm not taking anything away from anyone who has that challenge.
What AI Can't Read
Beyond those general concepts, I want to give you specific examples where I asked AI questions, and compare what it said to what I would say.
When I did this, I realized a couple of things. First, AI doesn't know the situation, and it doesn't hear the tenor of the voice. When I'm giving advice, it's not just what the person says. It's their voice. It's the question they don't ask, and how they ask the ones they do. It's having a husband and wife in front of me and seeing how they react to each other.
There's so much nonverbal information in addition to the direct verbal or written communication. Call it reading between the lines, and it would be impossible for AI to pick up. Much of the time, it's not the words but how they're said: the emotions, the eyes, how the clients look at each other, whether there's an issue with a child, a spouse, money, or taxes.
Set aside the comparison with AI for a moment. Even among human beings, this is probably what makes one doctor, accountant, financial advisor, or lawyer better than another: the ability to pick up this information. Call it emotional intelligence, or call it whatever you want. To me, it's a major factor in what makes one practitioner stronger than another.
Example 1: Trust Distribution Standards
I asked AI how to structure trust distribution standards. This is a classic question. A client is setting up a trust and wants to decide how, when, at what ages, and on what terms the children will receive funds after the client is gone. They want their children to be productive members of society, not simply to get money whenever they want.
I wasn't overly specific, and AI gave me the general information. There are three main types of distribution standards:
- Mandatory distributions.
- Discretionary distributions. These can be limited to a standard such as health, education, maintenance, or support, rather than left to whenever the trustee wants.
- Incentive or conditional distributions. Here the child receives money if, for example, the child does a certain thing or earns a certain amount.
When I dug a little further, I came across a joke I found very illustrative.
An older gentleman was planning to leave his retirement savings to his children and grandchildren. He asked AI, "How do I incentivize them? I want them to become productive members of society." AI asked him what was important to him. He said, "I have a lot of energy. I'm always on the go. I want my children moving, not sitting still."
AI said, "Why don't you put a monitor on each of your children? Your will or trust will say that whoever takes the most steps, whoever moves the most, gets the inheritance." The gentleman said, "Great idea."
The first child put the device on a fan. The fan went around and around, and he thought, "AI doesn't know what I'm doing. It's just getting feedback that I'm in constant motion." The second child, also a genius, put the monitor on a car and drove nonstop across the country. The third child had a wonderful dog. I'm also blessed with a dog, one who barks too much. He put the device on the dog, took it to the park, and let it run everywhere.
A month later, AI gave its verdict, and the inheritance went to the dog.
It's a joke about AI misunderstanding the goal. Rewarding the most active child sounded like a great idea, but the instruction wasn't clear, and the dog got the inheritance.
What would I do? We could spend a full-day seminar on this, because it's very specific. First, I typically don't like mandatory distributions. If there's a divorce, a creditor, or a business problem, why force assets out of a trust?
On the other hand, many clients want to incentivize their children, so we have to look at how much is in the trust and what the child's capabilities are. We have to look at all the factors. If you want to give the child some control, you can make the child a co-trustee, or give the child the power to hire and fire his or her co-trustee. When you're dealing with hundreds of millions or billions, though, that's not a great answer. If the child can reach everything, the child isn't incentivized.
So I can't tell you what I would recommend without the specifics. In general, I refrain from mandatory distributions, and I often don't like even the health, education, maintenance, and support standard, because circumstances change. What do those words really mean in practice?
To me, the most important thing is having the right trustees. When I'm alive, I know how to give money to my children, hopefully. Others may tell you I don't do it properly, but assuming I do, I want the right people in place after I'm gone. Some people don't have the right people, and they have to be more particular in the document. But it's very situational. If I have the right trustees for my children, I'm going to put my faith and trust in a person.
No person is above scrutiny, of course. Maybe I'll name two people who have to work together, plus a tiebreaker. AI won't get into that. It will tell you the standards, but not the checks and balances. In my opinion, it's always more about the people than the standards.
It's the same with estate litigation. People ask me how to protect against it, and I give them various answers. But the best answer is having family members who are virtuous, who aren't incentivized to fight, who are good people. As parents, we can only control so much, and a lot of that happens behind the scenes. AI couldn't possibly recognize any of it. Nothing it said about distribution standards addressed how much you trust your trustee. I'm not doing this topic justice, but I'm pointing out an issue or two.
Example 2: Trustee Compensation
Next I asked how much trustees should be compensated. AI gave me the statutory schedule in New York and the schedules elsewhere. It didn't address whether they should be compensated at all. There are too many factors for AI to answer that. You'd have to enter a thousand variables and look at the people's faces and circumstances. Will it make one sibling jealous if another is paid?
Then AI said something I completely disagree with: that you should explicitly specify trustee compensation in your document to prevent family conflict and avoid vague legal arguments. It got that from some journal.
With all respect to that journal, practitioners can disagree. AI gathers from many sources. Some are really good, and some are just okay. I've been doing this a long time. I don't know the journal AI drew from or who wrote the article, and I don't agree with it.
If you're naming a child or another family member as trustee or executor, I often tell clients to consider staying silent on compensation. Why stay silent, when the journal says never to? There are times when I want the child to be able to tell a sibling, "Look, New York law allows me to take compensation, but I'm not taking it." That child is now in better standing with the sibling, because he or she spent a lot of time on the estate and chose not to be paid. If the child decides to take compensation, the child is allowed to.
So there are circumstances where I'm deliberately silent. Would AI think of that? Would it see that if the child takes compensation, he's entitled to it, and if he doesn't, the other sibling will be happier and see that he acted in good faith? Every situation is different. I'm not saying never specify compensation, but there are certainly situations where you leave it to the executor or trustee to decide. There's no single rule.
Example 3: The Biggest Estate Planning Mistake
I asked AI for the top estate planning mistakes. It listed the simple things you hear from everyone: forgetting to fund the trust, failing to plan for incapacity, choosing the wrong trustee, and ignoring estate taxes.
You could read that in any textbook. Anyone could have said it. Maybe I wouldn't have said this 15 or 20 years ago, but after 30 years in practice, here is what I believe is the number one mistake: putting your children in a place of conflict. That means setting up an estate plan that could lead to a fight among the children.
For most parents, if not every parent, the greatest wish is that their children will get along. If they knew their children would fight after they're gone, many would tell you none of the planning was worthwhile. Some may reach an age where they can't fully grasp this, but if you reach them at the right time and place, no one wants their children to fight.
Will AI recognize that there's something deeper here than forgetting to fund a trust or planning for incapacity? AI isn't smart enough to know that. Or rather, it isn't emotionally intelligent enough. Many practitioners aren't either, because you have to go deep here, and AI is not deep.
Example 4: Sophisticated Planning Questions
I asked AI a number of other questions. In a couple of more sophisticated areas, such as international estate planning, foreign grantor trusts, and non-grantor trusts, it gave me a wrong answer.
I wrote back: "I disagree. You gave me a wrong answer." And AI replied: "You're right. I didn't think of that. These things are not a contradiction."
The moment I told it that it had gotten it wrong, it wrote back: "You're right, I got it wrong. I mixed up two concepts." So once you go deep enough, AI can't do it either. That may be solved over time, but at least for now, AI won't be able to help you that far in.
I'm not worried that AI will replace any experienced practitioner. Can it help with research and writing even now? Sure. But those are lower-level functions. As an experienced practitioner, I cut my teeth doing that lower-level work, so I don't know how the next generation will gain experience. But no one in estate planning needs to worry, and neither do people in comparable fields like psychology, accounting, or financial advising. There's far more to this work than producing knowledge.
Example 5: The Healthcare Proxy
There was another area. Unfortunately, I'm running out of time. The healthcare proxy is a simple document appointing someone to make medical decisions for you. I asked AI whom to appoint. It gave me basic answers, like a family member or a friend.
Then I asked, "What if that person has a different religious philosophy than you?" And AI wrote back, "You're right! That's something I should have said."
That's a big one. Often the person you choose can't make the decision you want, even if they'd like to, because their religious beliefs limit what they can and can't do. For many people, that religious or moral obligation comes before the obligation to carry out your wishes. AI didn't catch it.
A Word of Caution
As I hope I've illustrated, we're just not there yet with AI on these personal decisions. I don't think it will ever get there. AI is a robot. It doesn't have human feelings, emotions, or wisdom. Its intelligence is artificial.
I understand why clients run things through AI. But if they're working with an experienced practitioner, trust me, that practitioner knows everything AI will say. AI is like a textbook.
Clients sometimes bring me a tax technique with, say, 10 integrated steps. I tell them, "You're right, every one of those steps is fine on its own. But put them together, and it's what we call a step transaction." It starts to look like a plan to evade taxes. Avoiding taxes is fine, and evading them is not. AI couldn't grasp the step transaction concept: string too many steps together, and it looks as though you set out from the beginning to cheat the government. That's a judgment call AI can't make, and I don't think it ever will.
So my message to all of you is: use AI. It's wonderful for estate planning, but be careful. Many clients have wasted time and money coming to me with "I put your documents through AI and had the following question," and AI is always wrong. Once in a while there's a small point worth adding, but it ends up costing the client much more time and money. I understand that clients want to understand, and that's fine. Just be careful.
If you have an inexperienced practitioner, or you're unsure of their judgment, then using AI as a check and balance is great. But if you need AI to question your professional, maybe you chose the wrong professional. That doesn't mean people shouldn't educate themselves. But when I go to my cardiologist, I don't run the advice through AI to second-guess the doctor. I might ask a question or two for my own understanding. I don't think AI could know better than a doctor with many years of experience who has my specific information.
There's only so much I can feed into AI. Maybe we'll get there someday, but there are just too many factors. AI is only as good as the information you give it, and you can't give it all the information. It bases its judgment on books and records, and that's not how things work in real life.
You wouldn't go to a professor for business advice unless the professor was also a businessperson. For much of the advice we're talking about, whether medical, estate planning, tax, or financial, you want a practical, sharp, business-minded person. Professors are wonderful and teach a great deal, and many also have good business judgment. But that's the distinction.
I'm sorry I couldn't get more specific on many of these topics. Hopefully everyone got the gist of what I was trying to get across today.
Jonathan Shenkman: Thank you so much, Avi. Three quick items before I let you go.
First, my next webinar is Thursday, October 8th, at 8:30 a.m., on planning for state estate taxes, featuring Bruce Steiner of Kleinberg, Kaplan, Wolff & Cohen in New York City. I'll send out the invitation in the coming days. If you have a friend or colleague who would find these webinars of interest, they can subscribe at ParkBridgeWealth.com/webinars.
Second, you can follow my work on X and Instagram @JonathanOnMoney and connect with me on LinkedIn. You can listen to my weekly podcast, Jonathan on Money, on Apple, Spotify, or wherever you get your podcasts, and watch the practical planning videos I post several times a week on YouTube @JonathanOnMoney.
Third, please take 30 seconds to fill out the survey at the end of this program. It helps me improve these webinars and keep the content timely and interesting.
That concludes today's session. Please stay safe and healthy, and have a wonderful day, everybody.