AI Got the Numbers Right. So What Went Wrong?
An artificial-intelligence answer depends on the question and information it receives. Yet even a well-written question can produce an incomplete answer for a family. The tool does not independently know what money means to us, and often we have not yet found the words to explain it.
Recently, a prospective client came to speak with me carrying ten questions prepared by artificial intelligence. They were thoughtful, organized and seemed to cover everything. We discussed experience, investments, risk, fees and process. I answered each one.
In the final five minutes, after he put the list away, his real concern appeared. It was not in any of the questions. It was not a technical detail either. It was the issue that could change the entire decision. The tool had helped him interview us; it had not helped him recognize what was keeping him awake.
1. The question can be good and still be incomplete
The ten questions were good. That was not the problem. Their limit appeared because none reached the concern he had not yet expressed. We can ask the right question about the wrong subject, or provide every number while omitting the meaning that connects them. AI works with what we type; it does not live the conversation, notice a pause or independently recognize why we avoid a subject.
I saw this again in an AI-generated financial strategy. The recommendation made sense for the information provided. The client had asked about one part of his wealth and received an answer about that part. But he had not mentioned the business, real estate or pension. He was not hiding them. He simply did not realize those pieces would change the answer.
Once we included them, the liquidity, taxes, risk and sequence of decisions changed. Perhaps investing was not the first step. Perhaps a business obligation, tax date or future pension income needed attention first. AI had solved the written problem well. The written problem was not yet the real problem.
2. Money carries meaning that does not appear in the number
Consider a woman who receives an inheritance. She may ask, “How should I invest this money?” The answer may compare portfolios, returns and risk levels. But perhaps the money represents her parents’ lifetime of work. Perhaps she fears losing it, wants to help a child, cares for her mother or feels that spending it would be a betrayal. None of that appears automatically in a projection.
The same is true in a business sale. For a program, the result may be price minus taxes and expenses. For the person, it can also mean identity, freedom, guilt, security or the responsibility not to waste what the family built. A couple can receive the same amount and give it opposite meanings: one sees permission to enjoy; the other sees an obligation to protect.
Those meanings are not emotional decoration. They change concrete decisions: how much cash to keep, how much risk to accept, when to give money away, who should participate, what to protect first and what feels fair among children. Mathematics helps measure the options. It cannot decide what “enough,” “safe” or “fair” means to a family that has never discussed it.
3. Sometimes we need help discovering what to ask
Here is the most important limit: knowing how to write a good question is not enough. Sometimes we have the facts but still do not know which decision we are trying to make. Or we feel fear, guilt or family disagreement but cannot yet name it. An experienced professional does more than receive instructions. The professional listens for contradictions, asks what changed, identifies who depends on the money and helps uncover the question before trying to answer it.
That does not automatically make every human a better source. The difference lies in a professional process: knowing the family, verifying the facts, integrating taxes, estate matters and risks, and accepting responsibility for the analysis. Conversation does not replace calculation. It discovers which calculations matter, what consequences are missing and who will be affected.
AI has a valuable role: calculating, explaining concepts, comparing hypothetical scenarios and helping someone prepare for a meeting. But it should not direct a family financial plan on its own. Use approximate figures, protect personal data and check sources, dates and assumptions. Then comes the human question: What does this money mean, which family members will the decision affect and what have we not yet been able to say?
Conclusion
Artificial intelligence knows what money can do. It can project growth, calculate the cost of a loan or compare two alternatives. On its own, it does not know what that money means to you, your spouse, your children or your family.
That is why a family financial plan should not begin and end with an automated answer. First come the story, values, fears and people affected. Only then can technology serve us better: as a tool within a human decision, not as the one deciding for the family.
Frequently asked questions
Can AI create a personal financial plan?
It can calculate and model scenarios from the information provided. That is not the same as an integrated family plan: it does not independently uncover omitted facts, conflicting values or consequences the person does not yet know to ask about.
Is it safe to enter my financial information into an AI tool?
The prudent practice is to avoid sharing names, account numbers, identification numbers, tax returns or confidential documents. Use approximate figures or hypothetical situations and review each tool’s privacy policy.
How should I use AI before meeting a financial planner?
Ask it to explain concepts, compare hypothetical scenarios and prepare questions.
A question to take with you
Before accepting a financial answer, ask: What part of my life, my family or the meaning of this money is still missing from it?
Learn more: elaineking.com
About the author
Elaine King, CFP®, TEP, advises women, international families and family businesses. She has guided more than 1,200 families and is the author of eight books on financial education, financial planning and family governance.
Trusted sources
CFP Board: Code of Ethics and Standards of Conduct
FINRA: Regulatory Notice 24-09: Generative Artificial Intelligence
Investor.gov / SEC: Artificial Intelligence and Investment Fraud: Investor Alert
Federal Trade Commission: AI Companies: Uphold Your Privacy and Confidentiality Commitments
