Alpha AI

A new software program offers expertise in areas where most advisors don’t have expertise.

Every year at the T3 conference, speaker after speaker assures the audience that artificial intelligence is really important in the advisor space.  And then I look around and can’t find any functioning examples of this. 

Until this year, that is, when I was introduced to a program that quickly became the talk of the conference: fp alpha (https://fpalpha.com/), created by Andrew Altfest, President of Altfest Personal Wealth Management in New York City.  Unlike some of the grander visions of AI, fp alpha doesn’t talk to clients and give them advice directly.  Instead, it ‘reads’ client documents in more than two dozen aspects of their financial life, and then immediately generates observations and recommendations that the advisor can use to round out his or her planning advice in areas that many advisors are not currently versed in.

“fp alpha doesn’t overlap with the retirement planning or investment software providers,” Altfest explains.  “We wanted to build software that would address areas where the work today is manual, or where advisors are having to recollect something that they learned at a CE event four years ago.”

Let’s start from the beginning.  A client walks in the door, and you collect the client’s asset picture from your account aggregation software, along with the basic information like income, annual money being put aside for retirement and expected date when certain goals (like retirement) are anticipated.  You run this through eMoney, MoneyGuidePro or Right Capital, and you start managing the various investment accounts.

Now what?

That’s where fp alpha comes in.  The program has, currently, 18 different planning modules: advanced tax planning, basic tax planning, basic estate planning, advanced estate planning, home insurance, auto insurance, disability insurance, medical insurance, identity theft, student debt, long-term care, mortgage, life insurance, divorce planning, property tax, elder care, creditor protection and cross border planning.  (Two more modules are being developed as you read this: variable annuities and expense reduction.) 

You pick the modules where you want to provide detailed advice, and invite the client to take a short questionnaire on each of these topics.  At the same time, the client provides a document or two—the most recent 1040, the declaration pages from the client’s home insurance policy, total net worth and investment assets, the will and trust—and these are fed into the system.  The program can pull in client data from your CRM so some of the questions in the questionnaire are prepopulated.

Now comes the AI part.  Based on the information provided by the client, and the information pulled from the documents, fp alpha makes observations, some of it in the form of recommendations.  Since the modules all talk to each other, the data from one will inform another one, allowing for more specific recommendations or eliminating superfluous ones.  The estate planning output might note that the client’s heirs would, currently, be subject to inheritance taxes in her state, and she should either leave the assets to an exempt beneficiary (the spouse in most cases), or consider life insurance as a more efficient estate planning vehicle. 

The tax planning module, meanwhile, might recommend rolling IRA funds into a Roth, which would reduce future required minimum distributions and lower the client’s tax liability in retirement.

The questionnaire might reveal that your client owns millions of dollars worth of collectibles and art objects, yet the auto-scan of the home insurance policy doesn’t show any coverage for those valuable assets.  You have a very easy recommendation to make.

For each category, the program calculates a score, from 0-100, showing whether the client has optimized the recommendations or corrected the inefficiencies that fp alpha has uncovered.  There is also an overall score that pertains to all the different categories; a client with a score of “13,” who just walked into your office, has a lot of planning opportunities in front of her, while a client who has worked with you for years might have a score of “70” or “80.”

The interesting thing about these planning areas is how often some or all of them are neglected in a traditional planning engagement.  How many advisors thoroughly read through their client’s last will and testament, and compare it with their assets and the state estate tax rules?  Are capital gains allowed to be distributed for income tax reduction in the will?  Are contingent beneficiaries named in the will? 

Beyond that, do most advisors check their clients’ mortgage rate and balance, and continuously compare them to the most recent rates that are offered on the market?  How many advisors are able to model out the client’s current and future tax rates based on the current IRA balances, and show the value of partial Roth conversions?  “You can use the projection tool to see when you will hit a certain tax bracket,” says Altfest.  “Do you want to realize a certain amount of income today to keep yourself from hitting the highest bracket some years down the road?  Convert a certain amount, but not so much that you’re pushed into the next bracket, and see the long-term results.  Our software,” he adds, “will go from high-level planning opportunities down to the surgical, like how much you want to convert in a particular year.”

In short, the goal of fp alpha is to expand the definition of financial planning into areas of a client’s life where the profession has seldom gone before—and automate those important but esoteric areas where many advisors don’t have expertise.

“As you know, ‘financial planning’ is a very confusing descriptor,” Altfest explains.  “What is financial planning?  Right now, we have software around retirement, but all of these other areas, if I’m reviewing the client’s estate plan or the various insurances and trying to find gaps and opportunities, I’m doing this manually or I’m not doing it at all.  That,” he says, “is the problem that we’re trying to solve.”

You can see several pieces of one of the reports here, which happens to be the initial output from a fairly limited (early stage) client engagement.  The system will, of course, accept updates, which will determine how well the client and advisor have addressed financial inefficiencies in the client’s life. 

Moreover, as a way to help advisors communicate their value to clients, the program keeps a running total of the amount that the advice has saved them, based on the program’s initial advice calculations. 

“We quantify the value of the advice in dollar terms, and show the whole financial planning experience throughout the year in the task manager,” says Altfest.  “We wanted to put hard numbers behind the planning recommendations, which used to be vague and forgotten about when the focus was mostly on the portfolio.” 

In addition, if the advisor or client are confused about a recommendation, the program includes a detailed glossary that defines all the technical terminology and concepts that appear in the recommendations.

Will clients fill out the questionnaires themselves?  Altfest has been using fp alpha with his own clients for the past three years, and he’s found that some are willing to populate the information he needs online, while others need to come into the office and answer the questions together.  “At the very least, fp alpha allows you to be comprehensive in your meetings with clients, where you’re gathering more detailed facts and data,” he says.  “You have a guide to a greater number of issues to inquire about.”

Finally, where does the software’s expertise come from?  The programming behind fp alpha has a brute force quality to it; a team of more than 40 experts, from estate planning and elder law attorneys, CPAs, charitable planners, student debt experts, claims adjustors and insurance specialists in various areas looked at all the different possible outputs from Altfest’s own clientele, and envisioned other common and uncommon situations in their own experience, and then produced observations and recommendations that would be triggered whenever a certain circumstance would be encountered.  They’re also responsible for updating the information in light of new developments in their fields.

In that sense, the program isn’t doing so much independent thinking as it is matching data with prepopulated recommendations.  You can decide for yourself whether this is “true” AI, but the bottom line is that the program is likely smarter on a broader array of topics than most advisors, and it processes the information in the client documents almost instantaneously.

So far, in Altfest’s experience, no client has achieved a 100% score in fp alpha’s tough scoring system.  “By the time you get through the different areas of planning, the client will have more opportunities to improve the areas that you’ve already tackled,” he says.  The score is a moving target as a client’s life circumstances evolve in complexity and also, potentially, in different directions.  “There are always going to be things to do,” says Altfest.  “And on our end, we have a very rich product development path ahead of us.  There are a lot more cool things we’re working on.”