What AI Adoption Really Looked Like in 2025

AI for Advisors newsletter

Over the last month, I conducted a meta-analysis of 18 industry-related surveys from 2025, tracking how financial advisors are actually using AI.^* I’ve also trained more than 300 advisors through Horsesmouth’s AI education programs and, since 2023, spent more than 4,000 hours applying AI to specific advisor use cases.

What I’ve found is both reassuring and sobering: Nearly everyone is experimenting with AI, but very few are using it in ways that will create a lasting competitive advantage.

Here’s what the research shows about where the industry truly stands—and why it matters to your practice.

Observation #1: AI has crossed the awareness threshold

What the surveys show: Adoption has moved from curiosity to mainstream experimentation. Among retail advisors, 24% are daily users while only 22% abstain entirely. LPL’s conference survey found 78% of advisors either use or plan to use AI. On the institutional side, it’s even higher—78% of firms are implementing three to five generative AI use cases, and 73% of executives consider AI critical to their strategy.

What this means for your practice: You’re not behind. Almost nobody is “ahead” in any meaningful sense. The race hasn’t started yet because most participants are still figuring out which direction to run. The pressure you feel to “do something with AI” is real, but rushing into surface-level adoption won’t solve it. What matters now is depth, not speed.

Observation #2: AI adoption is shallow

What the surveys show: AI use is concentrated in low-risk, high-volume tasks. For retail advisors, the top applications are email drafting (43%), meeting notes (29%), and marketing content (24%). Younger advisors lean toward administrative tasks and research; older advisors favor communications. Investment process applications are emerging but cautious—product recommendations, plan generation, portfolio rebalancing—but they’re far from mainstream.

What this means for your practice: These aren’t bad use cases. Email drafting and meeting prep create real-time savings. But, they’re also the easiest applications to replicate. If you’re only using AI where everyone else is using AI, you’re not building advantage, you’re keeping pace. The advisors who will separate themselves are already moving toward the less obvious applications: client-specific scenario modeling, research synthesis for complex situations, strategic thinking support. That’s where the Advisor-AI Usage Ladder concept from my recent work becomes relevant—understanding when to use AI for speed versus when to use it for depth.

Observation #3: Why compliance feels like a roadblock

What the surveys show: Compliance obligations, privacy risks, and data accuracy dominate the obstacles list across all segments. Advisors cite compliance and privacy as leading barriers. Institutional surveys rank regulatory complexity and data accuracy as top concerns. Among risk management professionals, 71.6% identify data quality issues as their biggest challenge, with 35.3% struggling to reduce “false positives” also known as hallucinations, confident but incorrect AI answers to certain types of questions.

What this means for your practice: These are legitimate constraints that shape what’s possible in certain AI usage scenarios. But they’re also being used as reasons to avoid engagement entirely, which I feel is the wrong response. The right move is to work within these constraints deliberately. Start with use cases that don’t touch client data, investment recommendations, or detailed, unverified research. Master use cases outside of those areas build confidence. Then expand carefully into more sensitive applications as both your skills and the regulatory framework mature.

Observation #4: Policies and training are lagging behind usage

What the surveys show: Policies and oversight lag usage significantly and are seemingly contradictory. One survey found 78% of RIAs lack written AI usage policies, though another suggests progress with 82% now reporting their firms have policies. Among institutional firms, only 38% have established policies, but 64% have an AI committee or governance group. Firm training efforts remain nascent with one survey showing 44% of advisors feel confident using AI, and another showing 44% of firms are educating employees about it.

What this means for your practice: The firms and advisors who are setting up governance structures and investing in real training now are building the foundation everyone else will need eventually. If you’re waiting for your firm to “figure it out” before you engage, you’re giving away a 12–24-month head start to advisors who aren’t waiting. Individual advisors can make meaningful progress even before firm-wide policies exist. You just need to stay within appropriate boundaries while you learn. And if you haven’t configured something as basic as Custom Instructions in ChatGPT (which 68% of advisors haven’t), you’re leaving easy wins on the table.

Observation #5: Client trust is a moderating force

What the surveys show: Both advisors and consumers favor human oversight and are wary of autonomous decision-making. Client surveys indicate caution toward AI-guided investment decisions, while advisor surveys highlight high expectations for AI-driven portfolio recommendations. This creates an interesting tension between what advisors want AI to do and what clients will accept.

What this means for your practice: This is good news. Clients don’t want you replaced by AI, they want you enhanced by it. The sweet spot isn’t “AI makes the decision.” It’s “AI helps me think through more scenarios, spot patterns I’d miss, and communicate more clearly, so I make better decisions for you.” That framing resonates. It’s also honest. The advisors who will earn client trust around AI are the ones who position it as a thinking partner, not a replacement.

Observation #6: The real opportunity is client work, not admin tasks

What the surveys show: Firms are pouring AI investment into analytics, compliance automation, document management, and internal surveillance. But those tools mostly optimize the institution. There’s less evidence of investment in using AI to deepen client conversations, personalize planning scenarios, surface risks and trade-offs in plain language, and prepare advisors for more meaningful, higher-stakes meetings.

What this means for your practice: The next frontier is using AI for genuinely customized client work that doesn’t scale any other way. Imagine creating a personalized education piece that explains a complex strategy such as options collars or charitable trusts, or using examples from that specific client’s portfolio and life situation, not generic templates. Or imagine synthesizing research across multiple domains to support a complex estate-planning decision. These applications exist now, but they require comfort with deeper, more iterative use of AI. That’s the skill gap in action.

Observation #7: Time savings are real, revenue growth is TBD

What the surveys show: About 80% of advisors using AI report efficiency gains, and roughly half observe improved decision-making and customer experience. Sounds great. But many firms see only modest ROI and struggle to translate AI initiatives into sustained value. In private markets, 38% of respondents see no portfolio-level impact. Early gains are concentrated in time savings and workflow efficiency, but more complex outcomes like alpha generation or revenue growth remain speculative.

What this means for your practice: Here’s the reassuring part: You’re not missing out on some massive proven ROI that everyone else is capturing. The technology is ahead of our ability to measure its impact. Time savings are real and valuable, but translating those savings into revenue growth or client acquisition requires intentional strategy, not just better tools. The advisors who will show real ROI are the ones building AI fluency now, so when the clearer applications emerge, they’re ready to execute immediately.

The real issue is depth

Surveys can tell us who’s using AI and for what. But they can’t tell us the difference between an advisor who occasionally asks ChatGPT to clean up an email and an advisor who uses integrated AI into how they think, research, and solve problems. That’s the gap that’s widening.

The advisors who’ll separate themselves are the ones using it effectively and moving deliberately between quick tactical tasks and deeper strategic reasoning depending on what the situation requires.

A self-assessment framework

Answer these questions honestly. No one’s watching.

Surface-level use (Dabbler):

• I use AI mainly for writing and editing.
• I usually get what I need in one or two prompts.
• I haven’t customized any settings or created any templates.
• If AI disappeared tomorrow, I’d lose some time savings but my core work would be unchanged.
• I feel like I’m “supposed to use AI” more than I actually see the point.

Developing use (Competent):

• I use AI regularly for multiple business functions.
• I sometimes have longer back-and-forth conversations to refine outputs.
• I’ve configured basic settings like Custom Instructions.
• I’ve found 2–3 use cases where AI genuinely improved my outcomes, not just my speed.
• I’m starting to see where AI could support client work, not just my work.

Advanced use (Power User):

• I use AI as a thinking partner, not just a writing tool.
• I regularly engage in multi-turn conversations to reason through complex problems.
• I’ve built custom GPTs, saved prompts, or created repeatable workflows.
• I’m using AI for client-specific strategic work, not just general tasks.
• My team asks me for AI guidance because they see how I’m using it differently.

Where did you land? Here’s what to do next.

What to do based on where you are

Wherever you are on this spectrum, there’s a clear next step that will move the needle.

If you’re a Dabbler:

1. Stop feeling guilty. You’re exactly where most advisors are.
2. Spend 10 minutes configuring Custom Instructions. (If you need help, watch for an upcoming article.)
3. Pick one business challenge you’re facing this week and try a five-minute conversation with AI to think it through, not to get an answer, but to clarify your thinking.

If you’re Competent:

1. You’ve crossed the activation threshold. Now go deeper in one area rather than broader across many.
2. Identify your highest-value repeated task and build a reusable prompt template for it.
3. Start experimenting with AI for one piece of actual client work such as scenario analysis, research synthesis, planning memo preparation.

If you’re a Power User:

1. Document what you’re doing so you can teach it to your team.
2. Look for opportunities to move AI from “personal productivity tool” to “practice capability.”
3. Share your wins with others. We’re all figuring this out together, and the community learns faster when power users don’t keep their approaches secret.

The bottom line

The competitive advantage in AI comes from developing genuine skill while everyone else is still dabbling. The surveys show we’re in the early stages of a long transition. Most advisors are experimenting but very few are expert. That window won’t last forever. But it’s still open.

^*How I created the 2025 meta-analysis

I designed a deep-research prompt for ChatGPT’s Agent Mode to systematically locate every credible 2025 survey on financial advisor AI usage, extract the data, and synthesize it into a meta-analysis. I’m sharing this process to demonstrate both what the data reveals and how AI can serve as a disciplined analytical tool—not just a shortcut.

The surveys in the meta-analysis included: Horsesmouth: 2025 Advisor-AI Usage Survey; ISS Market Intelligence: Advisor Pulse 2025; LPL Financial: Focus 2025 Survey; Ernst & Young: GenAI in Wealth & Asset Management Survey; ThoughtLab & Grant Thornton: AI-Powered Investment Firm Survey; Accenture: North American Wealth Management Advisor Survey; Money Management Institute & Broadridge: 2025 Investment Advisory Pulse Survey; Advisor360°: 2025 Connected Wealth Report; Betterment: Advisor Solutions 2025 Advisor Survey; Fidelity Investments: AI Pulse Survey; Nasdaq & A-Team Group: Global Compliance Survey 2025; Financial Planning: AI Readiness Survey; ACA Group & Investment Adviser Association: Investment Management Compliance Testing Survey 2025; Natixis Investment Managers: Wealth-Industry Survey 2025; Charles Schwab: 2025 Independent Advisor Outlook Study; Dynamo Software: 2025 GP & LP AI Survey.

Ready to make the leap? Horsesmouth’s AI for Advisors Pro training programs provide the structured, advisor-specific approach that transforms occasional users into confident practitioners. Learn more at www.horsesmouth.com/aipro.