What Enterprise Data Leaders Can Teach Advisors About AI

Why Most Advisors Build AI Backwards

I've been watching advisors rush toward AI tools like kids in a candy store.

They download ChatGPT. They sign up for AI portfolio analyzers. They attend webinars promising instant transformation.

Then they wonder why the results feel hollow.

Henry Zelikovsky, CEO of Softlab360 - Engineering Ideas Since 2000 , sees this pattern often. After two decades helping enterprise firms integrate with a variety of data sources and implement machine-learning AI systems over the last 10 years, he's watching financial advisors make the same fundamental mistake.

They're building AI backwards, not recognizing the need for well-arranged data and not framing the context of their research

Why Advisors Fall Into the Conversation Trap

Many advisors start with simple conversational AI. They input a few prompts, see what responses they get, then try to figure out how to use the output.

Henry describes it perfectly: "It's like beginning a conversation with an unfamiliar person. You need to manage the dialog from a too broad topic to a more narrow and more relevant one."

But here's where it gets interesting.

The statistics reveal a striking contradiction. Over half of wealth management firms say AI tools are too complicated to use.

Yet 97% of financial advisors believe AI can help grow their business by more than 20%.

That gap between belief and execution -that's the backward approach in action.

Henry suggests treating AI conversations like client meetings – come prepared. Come with a stated goal, specific objectives, and an indication of what a fitting conclusion might look like.

His example: "I am profiling new Alternative Investment Asset Classes and offerings within them. I have been evaluating a few Alt products that could be suited for your portfolio composition. My view is that one or two of them would be a reasonable investment entry point into Alts for you, prospecting the next 5 years. What are your views?"

Notice the structure. Context. History. Specific timeframe. Clear question.

This is Digital Kaizen thinking in action. Small, deliberate improvements in how we structure our AI interactions create exponentially better outcomes.

How AI Becomes a Mirror for Advisor Thinking

Something unexpected happens when advisors start structuring their AI conversations this way.

They discover gaps in their own thinking.

Henry explains: "Advisors may not be fully aware of the level of knowledge and most recent information possessed by the client." They assume a near, say 2 to 5 years, time horizon makes sense. They assume clients understand alternative investments. Many new forms of Alts appeared in the last 2 years.

AI becomes a mirror that shows advisors where their reasoning breaks down.

This revelation changes everything. Suddenly, advisors realize they need more planning, more reasoning to interact with an AI tool. The more well-phrased the prompts are entered the more qualified the AI response will be

Advisors should take this under advisement to prepare for the AI tool conversation as one prepares for a discussion, with paragraph notations, tabular and chart samples, they discover they need deeper analysis than initially perceived.

The AI isn't just providing answers. It's revealing the questions advisors should have been asking all along. Learn from the AI responses to phrase your next prompt more precisely, more granularly. Take AI responses that are not on point as an indication that your prompt was not clear to the AI, revisit, rephrase, and improve. AI quality has been improving through ongoing machine learning of the provided data and its context. Learn to do the same – treat it as a continuous education and expansion of your knowledge.

Why Data Must Come Before AI Models

Here's where most advisors get it wrong. They jump straight to AI models without building the data foundation.

Henry's approach starts with data. "Lack of consistently accurate data is what we look to identify, evaluate, and offer software and machine-learning techniques to close that gap."

Traditional machine learning produces quality results after several training cycles, but only if you remove "statistical noise" from the data first. Once the results are consistently accurate from training models, we can reliably use them to generate insights from historical data and be ready to learn new data. Then you apply contextual AI using large language models to explain what you uncover.

For smaller advisory firms, this sounds like enterprise-level complexity. But Henry breaks it down differently.

"Practically, all RIAs subscribe to platform portfolio management systems and sometimes to more than one, using them separately for preferred features. To some degree, data in these platforms is relatively clean, but incompletions and discrepancies are common. May firms add their own data management rules to correct such cases?"

The key is to focus on data per use case. Don't try to fix all your data at once, and yet, when fixing data, be aware of all of its uses

Pick a specific direction your firm wants to take, then work on data relevant to that focus.

If you want to understand alternative investments, clean the data around alternatives. If you're focusing on retirement income strategies, start there.

This focused approach changes the entire conversation. Instead of saying "I want to understand alternatives," you say "I want to understand alternative investments that complement and increase ROI for specific clients with specific investment objectives over a relatively short time horizon, 5 years."

That level of specificity completely changes what data you need and what AI can actually deliver.

This is incremental improvement at its core. Start small, build systematically, then scale what works.

The SEC's AI Washing Crackdown: What Advisors Need to Know

The stakes for getting this right keep rising.

In March 2024, the SEC charged both former and current RIAs with making false and misleading statements about their use of artificial intelligence. Both cases involved firms promoting AI capabilities they didn't actually possess.

This "AI Washing" enforcement sends a clear message. Advisors cannot simply claim AI capabilities without proper infrastructure.

Here's what this means for your practice: document your AI processes, ensure your systems can explain their recommendations, and never market capabilities you don't actually have.

The regulatory environment demands AI systems that can explain their decision-making processes. This becomes particularly critical in high-risk use cases like investment recommendations.

Building AI backward doesn't just produce poor results. It creates compliance risks that can shut down your practice.

How AI-Informed Clients Are Changing the Game

The conversation around AI gets more complex when you consider client behavior.

Henry points out that advisors should "assume that clients are knowledgeable as well. Assume that they use AI personally, and they are prepared to challenge the advisor."

But then he adds nuance: "Clients, typically, do not use technology with historical, analytical exposure and AI on top of that. That's why they employ advisors to rationally explain market conditions and reference them when providing advice." AI is a reasonable tool to frame context in this preparation and form content to use in client discussions, but you need to work to make AI work for you at that level. To save time, you need to make time to put in time.

This creates two distinct client types. Self-directed clients who use complex technology become advisors to themselves.

But most clients want the human touch, conversation-level interaction with intelligent discussion.

For advisors working with pre-retirees and state employees, like I do, this human element becomes even more critical. These clients face complex decisions around pensions, Social Security, and retirement timing.

They need analysis, but they also need someone to walk through the implications with them.

AI enhances this relationship rather than replacing it. This is important – embrace AI, do not ignore it, and learn how to use it well.

Building Your AI-Human Feedback Loop

Henry describes the ultimate implementation as a complete learning system.

AI becomes the advisor's research assistant, helping find, qualify, process, and arrange discovery. The advisor absorbs this knowledge base to expand their vision.

Then they have a human conversation with the client.

But here's the crucial part: they capture that client discussion and feed it back into the data and AI process.

This creates a continuous loop where the human relationship actually strengthens the AI capability.

Each client conversation provides more data points. Each question reveals new areas for analysis.

Each objection highlights gaps in reasoning.

The system gets smarter because the relationships get deeper.

Your Step-by-Step AI Implementation Plan

If you're ready to build AI the right way, start with these steps:

First, pick one specific area of focus for your practice. Don't try to AI everything at once.

Second, audit the data you have related to that focus area. What's clean? What's missing? What comes from old systems that might have quality issues?

Third, practice structured conversations with AI tools. Set context. Define objectives. Ask specific questions.

Fourth, use AI responses to identify gaps in your own reasoning and presentation. What assumptions are you making? What supporting analysis do you need?

Fifth, create the supporting materials AI reveals you need. Charts, comparisons, historical analysis.

Finally, capture client conversations and feed insights back into your data and AI process.

This is Digital Kaizen methodology applied to AI adoption. Small steps, consistent progress, continuous improvement.

Creating Your AI-Enhanced Signature Value

Henry sees something bigger in this approach. "This becomes a signature value of such advisors. This value transcends family generations and may address an issue with transitioning generational wealth."

Think about that for a moment.

By 2027, AI-driven investment tools will become the primary source of advice for retail investors. Usage is projected to grow to around 80% by 2028.

Advisors who master the data-first, human-enhanced AI approach won't just survive this transition. They'll own it.

They'll become the bridge between technological capability and human wisdom. They'll help families navigate not just investment decisions, but the complex emotional and relational aspects of wealth transfer.

This signature value becomes something families rely on across generations.

The advisors who build AI backward will compete on price and convenience. The advisors who build it right will compete on wisdom and relationships.

After 22 years in financial services, I've seen technology waves come and go. The advisors who succeed don't just adopt new tools.

They integrate them thoughtfully into systems that grow stronger over time.

AI built on solid data foundations, enhanced by human insight, and strengthened by client relationships, creates something neither pure technology nor traditional advisory services can match.

Ready to take the first step?

Start with one client segment. Pick one data challenge. Build one structured AI conversation.

The Digital Kaizen for Financial Advisors series provides the roadmap for advisors ready to build AI the right way. Small improvements, systematic implementation, transformational results.

Your clients don't need another advisor with ChatGPT. They need someone who's mastered the art of AI-enhanced human wisdom.

That's not just a competitive advantage. That's a signature value worth building.

Chris Hensley is a financial advisor, podcast host, and creator of the upcoming book Digital Kaizen: Small Loops, Big Shifts in an AI World. With over two decades of experience guiding clients through complex financial decisions, Chris now blends his expertise in retirement planning with cutting-edge tools like AI, voice-first thinking, and behavioral science. Digital Kaizen is a philosophy for those who want to grow sustainably in a world that moves fast—combining human wisdom, technology, and tiny, honest loops of improvement. This article is a preview of the ideas explored in Digital Kaizen, due out later this year.👉 Want early access to tools and insights from Digital Kaizen? https://digital-kaizen-book.kit.com/6a16de43b8

Henry Zelikovsky is the founder and CEO of Softlab360, a New York–based software engineering firm (formerly Starpoint Solutions) that builds data- and AI-driven systems for financial services. He co-founded the company in 2000, guided it through a 2002 merger with Starpoint Solutions, and became CEO when it re-launched as an independent business in 2021. softlab360.com Earlier in his career, Zelikovsky held senior engineering roles at Credit Suisse, BNY/Mellon, and Merrill Lynch and founded SmartLink. He holds a B.S. in Computer Science from Brooklyn College and an M.S. in Computer Science/Software Engineering from NYU’s Polytechnic School of Engineering, and he frequently speaks on AI’s role in wealth management.

Grab the free Digital Kaizen Starter Guide and explore how to build better systems for thinking, learning, and working in an AI world.

#ArtificialIntelligence, #AIForAdvisors, #FinTech, #WealthManagement, #FinancialAdvisors, #DigitalKaizen, #AdvisorTech,