Blueliner Group
The question is worth taking seriously: if a capable AI assistant costs $20 a month and can explain AMT, model a Roth conversion, summarize an estate document, and answer a tax question at midnight, what exactly is a family office or personal CFO doing that justifies its cost?
It is a reasonable question. Anyone who uses AI regularly knows the tools are genuinely impressive — better than most early explanations you will get from a generalist advisor, faster than any reference book, and available at hours when no professional returns calls. The people asking this question are not wrong about what AI can do.
They are, however, frequently wrong about where wealth management actually fails.
What AI does well
Let us start with the honest acknowledgment: AI is a significant improvement over the previous default, which was either professional time priced for complexity or no easy answer at all.
Education and vocabulary. AI explains financial concepts accurately and at the right level of depth. If you want to understand how the alternative minimum tax works, what a qualified opportunity zone investment requires, or the difference between a revocable and irrevocable trust, AI is an excellent starting point. The field guide sitting elsewhere in these Insights — on equity compensation in San Diego biotech and life sciences — was written with the explicit goal of providing that kind of reference material. AI does this well.
Document summarization. Hand an AI a 40-page trust document, a partnership agreement, or a brokerage statement and ask it to identify the key provisions. It will do so accurately and quickly. This used to require a professional billing by the hour. It no longer does.
Scenario modeling on clean inputs. If you know your income, your tax rate, your basis, your holding period, and your planned transaction, AI can model the tax consequences accurately. It can run a Roth conversion analysis, estimate the AMT bite on an ISO exercise, or project the estate tax exposure on a straightforward balance sheet. On clean, complete, well-structured inputs, AI produces useful output.
Research and comparison. Evaluating whether a fee structure is reasonable, understanding how a particular investment structure works, or comparing two estate planning approaches — AI handles these comparisons well when the question is conceptual.
This is real value. For people with relatively simple financial situations — a single brokerage account, a straightforward tax situation, no equity compensation, no operating business, no estate complexity — AI plus a good CPA at tax time may genuinely be sufficient.
The problem is that people with $3 million to $20 million in wealth almost never have a simple financial situation.
Where AI falls short — and why the gap is larger than it appears
The structural limit AI advises on inputs you provide. It does not know what you have not told it.
AI advises on inputs you provide. It does not know what you have not told it.
This is the foundational limitation, and it is not a technical problem that will be solved by a better model. It is structural.
A personal CFO or family office advisor builds a complete picture of your financial life over time — not just the documents you upload or the questions you ask, but the things you do not know to mention. The ISO grant that was exercised three years ago with an informal agreement to extend the lockup. The trust that was drafted in a different state and has not been reviewed since a move. The investment in a friend’s company that may or may not qualify for QSBS treatment. The deferred compensation plan at a former employer that no one has reviewed in four years.
AI cannot find what it has not been shown. It does not notice the gap between what your estate plan says and what your beneficiary designations actually do. It does not cross-reference your corporate equity schedule to flag that a vesting event happened in October before the documents go to your CPA. It does not observe that the RIA managing your liquid assets has never spoken to the attorney who drafted your trust.
These are not exotic edge cases. In twenty years of working with founders, executives, and families with substantial financial complexity, the most consequential mistakes I have seen were not made because someone gave the wrong answer to a question they were asked. They were made because no one thought to ask the question.
AI does not coordinate.
This is the second structural limitation, and it is the one that matters most for clients in the $3M–$20M range.
A wealthy family typically has a CPA, an estate attorney, an investment advisor, perhaps a business attorney, and sometimes an insurance professional. Each of these specialists is competent. Each does their job correctly in isolation. The problem is that they do not talk to each other — and no one’s job description includes making sure they do.
The tax return is prepared based on what the CPA knows. The investment portfolio is managed based on what the RIA knows. The estate plan is drafted based on what the attorney knew at the time of drafting. Each of these professionals sees one piece. The cost of that fragmentation shows up quietly, in the seams: the tax deduction that was available but not taken because the investment advisor did not know the estate attorney had already moved those assets; the charitable contribution that was made after the stock sale instead of before, eliminating the opportunity to donate appreciated shares; the IRA beneficiary designation that names a deceased spouse because no one flagged the update.
What coordination actually is Coordination is not a knowledge problem. It is an accountability and continuity problem.
AI can answer a question about any one of these topics in isolation. It cannot observe the interaction between them, because it does not have persistent access to all of them simultaneously, and more importantly, it has no standing to initiate the conversation between the professionals involved.
Coordination is not a knowledge problem. It is an accountability and continuity problem. Someone has to hold the full picture, notice when the pieces are misaligned, and have the standing to call a CPA on a client’s behalf. That is not a task that a prompt accomplishes.
AI cannot act, follow up, or be responsible.
A conversation with an AI produces an output — a response, an analysis, a recommendation. What happens next depends entirely on you.
A personal CFO schedules the follow-up call, sends the document to the attorney, confirms the CPA received the vesting schedule before the extension deadline, follows up when the response has not arrived in three days, and notices when three weeks have passed and the estate update that was discussed still has not been executed.
The difference between knowing what to do and having it done is not a small difference at this level of complexity. It is frequently the entire difference.
The summary, in two columns
| What AI gets right | Where the seams fail |
|---|---|
| Handing you an accurate definition of AMT at midnight | Remembering to ask about the unmentioned out-of-state trust |
| Summarizing a 40-page partnership agreement | Ensuring the estate attorney and the CPA actually talk to each other |
| Modeling a clean, static Roth conversion | Managing the calendar and chasing the work until it’s done |
The question to actually ask
The question is not whether AI is better than a bad advisor. It clearly is.
The question is whether the financial complexity of your actual situation — the specific combination of equity compensation, concentrated stock, estate structure, business interests, tax position, and personal circumstances that you have — is one where the cost of the seams between your professionals is greater than the cost of someone to close them.
For a household with $500,000 in a Vanguard account and a W-2 income, probably not.
For a founder with $8 million in illiquid equity, a $1.2 million concentrated stock position from a prior acquisition, an estate plan last reviewed in 2019, a business generating $600,000 a year, and an upcoming liquidity event, the calculation is different. The seams in that situation are measured in seven figures over a ten-year horizon. The cost of coordination is not.
What this means practically
AI should be part of how you manage your financial life. Use it to understand your documents, test your assumptions, prepare for professional conversations, and stay informed. It is a genuine tool and a real improvement over the baseline.
Use it also to pressure-test the professionals you work with. If your advisor cannot explain their work more clearly than a well-prompted AI, that is useful information.
What AI does not replace is the function of someone who holds the whole picture, initiates the conversations that need to happen between your specialists, follows up when they do not, and is accountable for the outcome at the level of your actual financial life — not the question you thought to ask.
The $20 monthly subscription is real value. It is not a family office.
Blueliner Group is not a registered investment advisor and does not provide personalized tax or investment advice. The above is educational reference. Specific recommendations about your financial situation are made by your tax advisor, your registered investment advisor, and your estate attorney — ideally in coordination with one another.