What dreams have you had about reshuffling your financial life from top to bottom? How often have you thought that maybe you ought to try a computerized financial-planning service?

I have just done that, to give you an idea of what these services can -- and cannot -- do.

With the help of the New York accounting firm Goldstein Golub Kessler, I constructed a sample couple named Quinn, with two middle incomes, two children and some holes in the pocketbook. The fictitious Quinns tested four of the computerized services that are supposed to help you organize your money.

I used Merrill Lynch's Financial Pathfinder (price: $250), Aetna Life & Casualty's Personal Financial Analysis ($250), E. F. Hutton's Money Allocation Program ($175) and the Consumer Financial Institute's Money Management ($195). I filled out a lengthy questionnaire for each, and received a computer analysis of how the Quinns stand and what they should do next.

On the whole, the analyses made interesting reading (except Aetna's, which was incomprehensible), because they made me think about the Quinns' finances in a comprehensive way. For the most part, the suggestions were simple and sound, with some exceptions noted below. I'd say that the better services -- Merrill Lynch and CFI -- were worth the money. But they also raised some problems:

* Each service programmed its computers with different assumptions. Consequently, they didn't agree on the answers to some very basic questions. For example, based on exactly the same information, E. F. Hutton concluded that the Quinns needed $421,000 worth of life insurance; CFI advised $127,000; Merrill Lynch -- the only service that thought life insurance was needed on the wife -- suggested $66,000; Aetna, surprisingly, said the Quinns needed no life insurance at all.

(Aetna's no-insurance finding probably would change, once a live agent became involved in the scenario. Aetna's John O'Connell told me that I had probably underestimated the couple's need for insurance when I filled in the questionnaire. "Normally, these questionnaires are filled out with the help of an insurance agent," he said, "and that's the sort of thing they would watch for." So the person who helps you fill out the forms -- usually a stockbroker or insurance agent -- can influence the computer analysis just by guiding the numbers you put down.)

* It's easy to err when filling in the questionnaires, because the directions aren't always clear. And when you err, the computer will, too. I misunderstood a CFI question, and as a result, CFI seriously underestimated the amount the Quinns should save for retirement.

One important point is that I did not notice this error myself. Accountant Stuart Kessler, who reviewed the results, had to point it out to me. Had I been an average client, and taken the computer's conclusion as gospel, I'd have been up a creek.

* Sometimes the plans seemed to be selling products rather than giving disinterested advice. E. F. Hutton, for example, recommended that the Quinns spend $1,873 a year on a universal-life insurance policy. Because of that policy's high cost, they would not be able to afford as much total life insurance as they needed, nor would they have enough money left to buy sufficient disability insurance or start a savings plan.

But they could have met all of their goals had Hutton recommended a low-cost term-insurance policy instead. Maynard Engel, who runs the Hutton service, conceded that the universal-life suggestion was "an inappropriate recommendation," and sent me a replacement plan based on term insurance for $728. Engel says it was all a mistake; he examined other plans, he said, and I got "the one out of 100 that went wrong."

But had you gotten such a plan, you might have gone ahead and bought the expensive and inappropriate policy.

Another thing about Hutton: It went for broke on some of its proposed investment returns. If you choose high-risk investments, it suggested, you could earn 21 percent a year after inflation! With what? Specialty stocks, commodities and covered call options, Engel says. I sure would hate to stake much cash on that strategy.

* The Aetna plan was impossible to read or understand without the help of an Aetna agent. Hutton also had some confusing charts. The Merrill and CFI plans were easier, but still unclear in spots to someone unfamiliar with financial plans. In the end, you will probably need help from a human financial planner, and therein lies even more traps than you encounter with machines.