Six months after predicting that AI would transform wealth management, a practice principal reports that the anticipated 70% reduction in email time has not materialized. Instead, the most significant gains have come in the quality of advice on complex tasks, not in hours saved. The principal, who runs a registered investment advisory (RIA) and has been testing AI tools since early this year, shared the update in a detailed account of what has and hasn't worked.
One example involved a retired client with a concentrated position in his former employer's stock, carrying substantial unrealized gains. The client wanted a multi-year divestment forecast. Using Anthropic's Claude system, the advisor produced a comprehensive analysis in about an hour of prompting. Doing it manually would have taken at least 15 hours and would have been less thorough, less data-rich, and less polished. The AI didn't compress existing work; it enabled better work that wouldn't have been feasible otherwise.
Trust and the email agent
The email agent, which drafts responses in the advisor's voice, is about 80% accurate but still requires review. The advisor doesn't fully trust it yet, so every draft is checked. Where it helps is in repetition: a thoughtful answer to an investment question that might take 20 minutes from scratch can be adapted from a similar prior response in a minute or two. The advisor expects to reach the 70% time savings once confidence grows, but complex matters will always get human review. This is an experience-based business, and the advisor doesn't want it to become an AI commodity business.
One unexpected benefit: the agent sometimes raises points the advisor might have missed, such as wishing a client well after a recent surgery. The in-house AI specialist calls this "perfect recall," a feature that adds a personal touch.
The unglamorous work: data and compliance
The practice has set up a separate entity, Destiny Intelligence, to offer its AI process to other RIAs. A six-month proof of concept was planned, but the rollout to other firms likely won't happen until early next year. The advisor wants to be close to 100% satisfied before putting his name behind it. Two hurdles explain the delay: data integration and compliance.
Data is the biggest obstacle. Most RIAs struggle to connect systems like Salesforce, Orion, and eMoney, each of which has added AI features, but the data must be cleaned first. Compliance is a heavier lift than expected because AI is still new in the advisory world. The practice uses the most protected version available, but nothing is 100% protected. They are also adding AI cybersecurity insurance, which is distinct from a standard cyber policy. Every firm should ask its carrier about the difference.
Governance and competition
The advisor acknowledges that some dire warnings about AI reflect real concerns, but believes they are further down the road. He advocates for self-governance, with AI companies overseeing one another, rather than heavy government involvement. The U.S. leads in AI, with China close behind, and he argues against surrendering that lead. The debate was on display at the All-In Summit in Los Angeles in September, when President Donald Trump phoned Nvidia CEO Jensen Huang onstage. Awareness of the downside is smart, but becoming uncompetitive is not.
AI in private markets
AI also powers Destiny Access, a platform for qualified purchasers investing in direct deals and alternative assets. Direct investments are notorious for reporting lags. Through AI feeds, the platform updates deal valuations as soon as information is available, though not client valuations due to liability. For example, when a company like Crusoe completes a new raise, the platform reflects it on day one, and everything is still screened by humans.
The advisor's digital twin sits on both the public and private sides of the site. The public agent answers general questions and escalates complex ones to the advisor. The private agent handles holding-specific questions. Anyone who wants a person still gets one.
If the advisor could tell himself one thing, it would be to start earlier. He estimates that 95 out of 100 advisors say they use AI, but only five say it delivers significant benefit. His practice reached that group in about six months, largely because they hired a specialist rather than trying to do it alone. With clients now using AI to challenge financial advice, that gap will only matter more. For advisors looking to improve their practice, the lessons from this experience are clear: focus on data quality, compliance, and trust-building before expecting time savings.
