Financial advisors are finding that free artificial intelligence tools often fall short, not only in features, but also in the privacy protections required to handle sensitive client data.
Data entered into free and even many paid consumer AI tools is often used for AI model training. Sometimes users can opt out of their content being used to train AI, but not always.
While premium versions generally provide better features and services, the critical difference is that they also offer a higher level of protection
Financial advisors should carefully assess which AI tools support their compliance, confidentiality and ethical obligations, said Marta Hovanesian, a privacy and data attorney at
"Free versions do not provide the appropriate safeguards needed, even if you opt out of AI model training," she said. "Consider matters such as data retention, security and access controls."
READ MORE:
How advisors choose their AI tools and features
George Chang, financial advisor and principal at
Chang said he has tried the Google Meet note-taker, Wealthbox AI and Jump. While he found Jump to have more features, he didn't feel it was worth the extra cost (and he didn't like having yet another application to open). Instead, he went with the note-taker from WealthBox, which he was already using for his CRM.
Chang also uses several other AI tools, including Perplexity. Upgrading to Perplexity's paid version was well worth the cost, said Chang.
"In addition to Perplexity's strength of researched, sourced answers, the paid version allows me to set up what they call 'Spaces,' or contexts, such as compliance, marketing and tax, which I can use for particular questions," he said.
READ MORE:
Reggie Fairchild, financial advisor and president of Mount Pleasant, South Carolina-based wealth management firm
"It helps us turn financial jargon into real-life conversations our clients actually understand," he said.
The firm is also exploring several other tools — Claude for its responses, Gemini for integration with Google and Grok for news and headlines, said Fairchild. Each has its strengths, but for now, ChatGPT is his favorite, he said.
"It's like chatting with a sharp junior planner who's always ready for more," he said.
His firm started with the free versions, but it didn't take long to see the limitations.
"It's like serving sweet tea without ice — fine in a pinch, but not how we do things in South Carolina," he said. Leveling up to the paid versions of his AI tools was a "small investment that gives us a big lift — especially in content planning and improving how we communicate financial strategies."
"It's like hiring a part-time writer who works fast, takes direction, and never needs reminding what your voice sounds like," Fairchild said.
Privacy controls often lacking outside of enterprise editions
The upgrade to paid tools gave Fairchild's firm faster, more accurate responses — and most importantly for him, stronger privacy controls.
"As a fiduciary firm, that matters," he said. "We expect that future versions of AI will anonymize personally identifiable information before it ever leaves our local computers. That will further protect our clients."
But while paid consumer versions offer great value for research, Chang said he feels none meet the strict privacy needs for sensitive client data.
For example, both Microsoft Copilot and ChatGPT offer advanced data protection to enterprise customers that is not available to even paid non-enterprise users.
Chang said he also uses Gemini, which is contained within Google Workspace, and his note-taker, which has clear privacy and data security agreements.
"True security requires enterprise subscriptions, and even then advisors need to read the agreements carefully," he said.
This is where larger firms have the advantage of being able to bring AI engines in-house and connect to their data in a closed ecosystem, said Chang, though that's not the case for him.
"Smaller RIAs like myself currently must take a piecemeal approach, using public AI tools for knowledge, but strictly avoiding free tools to analyze sensitive client information," he said.
For now, Fairchild said his firm's philosophy is to pay for professional services and take additional manual steps to protect clients.
"I wouldn't leave client notes out on the porch," he said. "Same goes for anything we feed into AI."
Client trust is the cornerstone of his work, said Fairchild. The paid version of ChatGPT allows him to use AI responsibly — "without crossing lines that matter."
"Trust is hard to earn and easy to lose," he said. "We'd rather stay ahead of the curve than clean up a mess."
