How to Evaluate WealthTech AI for Financial Advisors

It seems like everywhere you turn these days someone is mentioning AI. Within the WealthTech world, as of 2026, there are already 100+ firms on The Oasis Group’s AI WealthTech Map, featuring “AI-first firms solving wealth technology problems.” Plus, every month there seems to be an announcement about an existing platform incorporating a new AI feature, such as RightCapital’s Smart Import, introduced in March 2026 and RightCapital's Iris, introduced in June 2026 (after this post was published).

So, questions abound. How can financial advisors use AI productively and safely? Which AI systems are best for your processes? How much AI is too much AI? For these reasons, our Q1 RightInsights event was focused on key WealthTech and AI trends. We hosted industry experts Michael Kitces, Chief Financial Planning Nerd at Kitces.com, for “Financial Planning Tech and AI Trends for 2026”, and John O’Connell, CEO and Founder of The Oasis Group, for “What AI Can (And Can’t) Do for Your Firm”. Here, we collect their thoughts and answer some practical questions:

How can you determine which AI WealthTech will work for you?

It’s easy to be impressed by each piece of AI WealthTech that you evaluate, especially if you haven’t yet seen the practical value of AI in real life. John O’Connell called it out directly, “Every one of these tools looks impressive...Do not go into a demo with rose-colored glasses.”

So that you don’t sign up for every service that excites you, O’Connell recommends walking into or logging onto a demo with a detailed list of your own firm’s requirements. Make sure you ask questions to see examples within the software that address your specific use cases and challenges. If you start this way instead of potentially being distracted by flashy features you may not use, O’Connell noted, “you’re going to find that the number of tools that you have available to you, that universe, gets compressed relatively quickly.” Knowing exactly what you want will help narrow down the choices.

What are the warning signs you’re buying “shiny objects” instead of solving problems?

If you are constantly switching tabs, re-entering data, or juggling logins, you may be past the point where technology helps.

Michael Kitces described the “negative productivity curve” in very human terms: “There comes a point where we've got so much tech that even if all the things are awesome, I have to spend so much time figuring out how to learn and use and weave together and integrate all of those things into my workflows. If your head hurts because of how many software packages you have to switch between, you might actually be past the point where it's helping.”

This is also an argument for keeping your “anchor tools” (identified by Kitces as financial planning software, CRM, and portfolio management) as the center of gravity, and prioritizing AI that fits naturally into those workflows. The best AI reduces manual work and follow-up friction without creating new systems someone has to manage.

How do you assess data accuracy (so AI doesn’t just scale mistakes)?

When you generate something with a large language model, I'm a huge fan of having what's called a human in the loop.

A warning from O’Connell belongs on every RIA’s AI checklist, “AI tools are only as good as the data that they capture. If they don't capture great data, garbage in, garbage out. They're just going to do it a lot faster.”

Accuracy can become especially tricky with finance-specific language. O’Connell used “529” as a concrete example: “In many cases, what we find is the generic notetakers don't capture that very well. In fact, they may capture it as an account number, 529, as opposed to the account type of a 529.”

When using AI for another popular use case of client communication, which can include meeting prep, emails, or generating blog posts, O’Connell said, “When you generate something with a large language model, I'm a huge fan of having what's called a human in the loop.”

In discussing hallucination (when AI makes up data to support your decision) and bias risk (when AI platforms only reflect the data they were trained on), O’Connell recommended asking for citations within your prompts. "Understanding what hallucinations are, how to identify them, and how to combat them, for example, by asking for referenceable sources for data, is very important for you to be able to reduce your hallucination risk or bias risk.”

When it comes to matching your firm’s tone and position, O’Connell suggests including guidelines for voice and structure when prompting. “You are accountable for what is generated by your large language model. You need to review that to make sure that what is generated matches the fiduciary responsibility you have as an RIA and your firm's tone.”

Why is “document extraction” an AI sweet spot?

Kitces explained the practical strength of today’s LLM-based AI, “It's good at synthesizing large volumes of words and doing things with language.” He then pointed to a particularly high-value use case, “All things where AI extracts information from documents.”

He described the exact workflow advisory firms want, “Here are 27 pages of PDFs of my client stuff. Read it, figure out the numbers that matter, and put them into my planning software, or my CRM, or my other software system, so I don't have to do data entry. It’s a great sweet spot.”

Kitces noted that there may be another down-the-line benefit of data extraction tools. “You're going to help my client get to more actionable recommendations that actually improve their financial future, which means I can go deeper with them. I can charge full value commensurate to the service, which means I've got a healthier business with better productivity, more margin to reinvest, and better word of mouth for growth, because I'm actually helping my clients do more awesome things.” Document extraction tools can save time and allow for deeper planning with better recommendations.

If you’re evaluating extraction tools, you may be interested in RightCapital’s Smart Import. The tool reads financial documents such as statements and meeting transcripts and brings relevant data into a financial plan, reducing manual data entry and speeding up the whole process.

What is the benefit of AI notetakers?

The most highly adopted category of AI WealthTech in the industry thus far has been notetaker platforms for meeting documentation: standalone tools as well as capabilities within platforms you’re already using.

O’Connell and Kitces both lauded the time savings AI notetakers create for advisors, with O’Connell adding, “You can stay present in the conversation knowing that the notetaker is actually taking really good notes.” and Kitces calling them, “A great piece of tech for a very concrete problem.”

How can advisors use AI for compliance review?

Both Kitces and O’Connell touched upon the use of AI in compliance review during their respective RightInsights sessions. They agreed upon the specific use case of AI reading every piece of correspondence in and out of the firm.

Kitces noted, “We don't have to look for keywords like guarantee and fraud and complaint, because it understands natural language enough to spot questionable sentiment concerning statements being expressed by the client, worrisome notes, that might surface fraud.”

O’Connell called out this capability as a huge opportunity for the compliance team, “Your compliance team will be much more efficient in being able to review your documents.”

Can AI replace financial advisors?

In short, no. Per Kitces, "Businesses that are built around human relationships first and foremost don’t move at the speed of tech. They move at the speed of relationship, and the speed of relationship is slower.”

Many of the AI advances in the market so far have been geared toward the DIY, the do-it-yourselfers, who are not speaking to advisors anyway. Kitces continued, “I've never lost a client because they were like, ‘I can do all this on Mint.com or Monarch Money, I don't really need an advisor anymore.’ Because delegators don't want to do that! They delegate!"

If you don’t use AI, are you toast?

According to Kitces, “No client up and leaves because they're like, ‘I talked to this other advisor, and he has a cool piece of software. That's not why clients leave. We either have good communication relationships with them, or we don't, and if we don't, they look elsewhere. And so, in that world, you can't be fast disrupted.”

Kitces also believes that in some situations, it may make more sense to hire a person to perform a specific task than to pay increasing software costs for however many advisors are in your firm. Software is not just a one-time cost, as O’Connell also touched upon. Costs such as time spent training, migrating data, troubleshooting, as well as if there are additional fees for storing or exporting data if the relationship ends should all be considered.