SHANNON ROSIC:
Hi, everyone, and welcome to this Wealth Management Fast Chat. I'm Shannon Rosic, director of content solutions, and I'm thrilled to be joined by Cindy Griffin, vertical marketing specialist for financial services at Smart Communications.
Cindy, great to see you.
CINDY GRIFFIN:
Good to see you too, Shannon.
SHANNON ROSIC:
Well, I'm excited to jump into this conversation because I know we're going to be exploring where AI can actually meaningfully improve onboarding, where human involvement still matters, and what it takes to connect all those pieces into a more seamless experience.
What can AI be trusted with in the onboarding process, and where do we still need a human in the loop?
CINDY GRIFFIN:
I think we're seeing a significant shift in what AI can do in wealth management. We're moving beyond AI that simply summarizes information—which was among the first tools—or answers questions with generative AI.
Newer AI tools are able to work across existing technology platforms and initiate activities, and they're potentially going to be able to execute increasingly complex tasks. I think that opens up tons of exciting possibilities for onboarding, but I think there are two very important considerations.
The first is that wealth management firms generally have highly specialized and often customized onboarding processes that are part of their value proposition. A general-purpose AI platform may be able to coordinate activities across systems, but those activities are still going to need to follow the firm's specific business rules, the regulatory requirements they operate under, and their approval processes.
The second consideration is that AI still needs reliable information to work with. That includes information the firm already holds and new information provided by the client. If that data is incomplete, inaccurate, or inconsistent, even a sophisticated AI tool may struggle with how it moves the process forward reliably.
I think that's where the human-oversight part becomes really important. We need to distinguish between allowing it to perform a task and giving it the authority to make a decision.
AI might identify an inconsistency or maybe even prepare a suitability assessment for review, but the firm still needs to exercise judgment. An advisor or compliance professional still needs to come into the process, and the firm still needs to retain accountability there.
The goal is really—I mean, we have to use AI, right?—but we need to give AI the freedom to operate within well-defined controls, supported by trusted data and firm-specific processes, while preserving human accountability where it matters most.
SHANNON ROSIC:
Look, the reality is that "human in the loop" sounds pretty straightforward until you consider just how many people can actually be involved in onboarding, especially when firms are trying to do this at scale.
If human involvement is required, how do you effectively engage clients, advisors, client service teams, and compliance throughout a process that needs to scale?
CINDY GRIFFIN:
I think the answer is not to insert a person into every single step. That absolutely defeats the purpose of automation, AI, or anything that you want to do.
At scale, you want to design a process around exceptions. Most straightforward cases should move forward automatically, while the right person is brought in only when something requires review, approval, or judgment.
For example, a client might complete a guided digital onboarding journey, with data validated as it's entered—assuming the firm has designed the process so that it's intuitive and adaptive, surfacing only the appropriate questions based on the client's circumstances.
If everything meets predefined rules, the process should just automatically continue. Think about that as straight-through processing for client onboarding.
If something falls outside those rules, though, that case can be routed to an advisor, client service team, or compliance officer, with the relevant context already assembled.
That's really how you're going to scale: not by adding more people to the process, but by making sure that their attention is focused where it actually adds value.
SHANNON ROSIC:
To me, this all sounds a bit like putting together a chain of events, right? A chain is only as strong as its weakest link, and AI still doesn't solve every part of the onboarding process, to your point.
What are some of those critical links in that chain that AI is not yet adequately addressing?
CINDY GRIFFIN:
I think one of the most important links in this whole AI story is the information itself. AI tools are increasingly capable of accessing information across multiple systems, but having access to information doesn't necessarily mean that the information is accurate, complete, or even consistent.
Think about a client's complete financial profile. Their information might exist in multiple systems, like a CRM, a financial-planning platform, or even just several PDFs and documents. What happens if those sources disagree? Which information should AI rely on?
That's why I think we need to look much further upstream, to where the client information enters the organization in the first place. Firms need reliable ways to capture and validate information, structure it appropriately, and make it available to all the systems that the process depends on.
That isn't just a one-time exercise with every new client. Every new account change or service request introduces additional information that needs to be accurate and usable, right?
I think another critical link is the client experience itself. Even if AI can coordinate activities behind the scenes, clients still need an intuitive, easy way to provide information, complete documents, and take all the required actions to give you that information.
The firm still needs to ensure that the correct disclosures are produced, that appropriate approvals are obtained, and that there's an auditable record of what happened.
AI can increasingly work behind the scenes, connecting parts of this journey, but firms are still going to need the trusted data, governed communications, and connected client interactions that allow the journey to work from beginning to end.
SHANNON ROSIC:
Clearly, the answer isn't putting AI everywhere. It seems to be more that firms need to be much more intentional about where they're actually deploying it.
Where are the best places to insert AI into a broader onboarding process? And conversely, where does it just not belong?
CINDY GRIFFIN:
I'd start by looking at the entire onboarding journey and look for where AI could be useful to interpret information or provide context to improve the experience. And when I say "experience," that's for either the clients or the staff.
Good examples of this include conversational guidance; helping clients or advisors understand complex questions; extracting data from uploaded documents; summarizing information for an advisor; or assisting employees who need to understand where a case stands, right?
Where I would be more conservative is in things that are compliance-heavy: final compliance decisions, suitability determinations, or any step where the firm needs to be able to clearly, clearly explain why a decision was made—basically, anywhere human judgment and accountability are essential.
Importantly, not every step needs AI. If rules-based validation can determine that, for example, a Social Security number has the wrong number of digits, just use a rule. If a workflow can automatically route an application for approval, use that workflow.
I've used this analogy before regarding AI: the old saying, "When you have a hammer, everything looks like a nail." AI is the hammer today. The goal should be to use AI where the intelligence that it brings to the table adds value, but not just because AI can be inserted there.
SHANNON ROSIC:
There's also an operational question that comes with all of this, because implementing AI is one thing, but somebody has to own and maintain these processes over time.
Can AI be maintained by the same people doing this work today, or does this new AI paradigm require more specialized resources?
CINDY GRIFFIN:
I think the answer is both. You're going to need specialized expertise around AI governance, model security, and risk. But the people who understand the onboarding process today still have to be deeply involved.
Operations and client service teams know where exceptions happen. Advisors know where clients get confused. Compliance knows where judgment and controls matter. Those people should help define how the AI is used and, just as importantly, where it's not used.
The goal should not be to create an environment that only a small group of people can understand or change. Business users really need to retain ownership of the process, the content, the rules, and ultimately the client experience.
I think about this as a shared operating model. At Smart Communications, we really believe it's important that business teams continue to own things like the content, rules, and experience that they're responsible for today, while technology and risk teams provide the specialized oversight needed for AI.
SHANNON ROSIC:
Let's bring this all together with the end goal that firms have been working toward for years, which is creating a more seamless, straight-through onboarding experience.
Can AI realistically create a straight-through onboarding process, or are there still too many gaps and pauses that could ultimately lead to client abandonment?
CINDY GRIFFIN:
I think AI can help us get much closer to straight-through onboarding, but I don't think AI alone is going to create it.
The main reason is that the biggest causes of abandonment in the onboarding process are often much more mundane: you're asking clients for information twice; sometimes questions are just unclear; you have missing documents; there are manual handoffs and waits for approvals; or there's no visibility into what happens next.
The strongest onboarding experience for your clients would combine several things: intelligent guidance where it helps, structured data collection, real-time validation, automated workflow, digital signatures, personalized communications, and clear updates.
For simpler clients, I think AI can help create a highly automated experience. For more complex high-net-worth and ultra-high-net-worth relationships, there are still going to be pauses and human involvement, but those pauses are going to feel more intentional rather than accidental.
I think that's really the distinction. The goal is not zero human involvement; it's zero unnecessary friction. That's where we think the real opportunity lies.
We're combining capable AI with reliable data collection, governed processes, and connected client experiences so that onboarding can continue moving forward even when human intervention is required.
SHANNON ROSIC:
Well, Cindy, this has been a great conversation. As always, I appreciate you joining me and sharing your perspective on where AI fits into onboarding—and, just as importantly, maybe where it doesn't.
CINDY GRIFFIN:
Great. It was great to be here.
SHANNON ROSIC:
Thanks, everyone, for watching, and we'll see you next time.
Explore research, expert insights, and resources related to this conversation.
Smart Communications™ is the only provider of a customer conversations management platform. Built from the cloud up, we help enterprises deliver SMARTER conversations across the entire customer lifecycle – and at scale – and empower them to succeed in today’s digital-focused, customer-driven world while also simplifying processes and operating more efficiently. This is what it means to scale the conversation.