Zehra Cataltepe is the CEO of TAZI.AI, an adaptive, explainable AI and GenAI platform for business users. She has 100+ AI papers & patents.
For wealth management firms, finding and converting the best potential clients is essential for organic growth. Historically, most firms have relied on referrals, but cold outreach and digital marketing are becoming increasingly popular growth tactics. Unfortunately, these channels haven’t yet been delivering the expected results given the ever increasing competitive market. This is mainly because they use insufficient numbers of signals to determine a fit.
On the other hand, AI-based lead scoring models can predict with higher accuracy which prospects are most likely to convert into clients and help wealth management organizations focus on the best opportunities. However, even with AI, most AI models are black boxes, which means that they generate a list of leads but don’t explain why certain prospects are ranked higher than others. I have heard numerous times how black box AI leads weren’t really usable in wealth management firms and how, as a result, firms are losing trust in AI.
The challenge with black box AI lead scoring models is that they generate predictions, but they don’t explain why a lead was ranked high or low, this leads to:
• Lack Of Trust: Firms can’t see why AI chose certain leads and can’t rely on recommendations.
• Generic Outreach: Without understanding AI scores, marketing teams and advisors can’t personalize messaging to increase the conversion rate.
• Missed Opportunities: If AI ignores important client traits, valuable leads could be overlooked.
On the other hand, with explainable AI (XAI), wealth managers can understand why leads are generated and they can adjust their strategies more effectively.
While there are many methods in XAI, such as variable importance methods (Shapley, LIME) and counterfactuals (DiCE), in this article, we will concentrate on human interpretable surrogate models, such as simple decision trees, or linear models, to help firms see why a lead is likely to convert and make better marketing and outreach decisions.
Here is how you can do this in 3 steps:
Step 1: Create An AI Model To Predict Your Ideal Client Profile (ICP)
Before investing in client acquisition to grow organically, you should first understand your most valuable existing clients.
A customer churn prediction model helps identify who will stay with your firm the longest and will bring in the most revenue. In order to determine ideal customers, you would also use AI models to predict advisory fee contributions, and to predict which service (investment management, estate planning, tax strategy) each client would benefit from the most.
You would segment (group, cluster) your clients based on the predictions of the attrition, advisory fee and service models and use explainable AI to identify your ideal client profile (ICP). The insights you get from AI explanations could be, for example:
• Clients with $800,000+ in assets and a strong interest in both tax planning and investment management tend to stay the longest and generate high advisory fees.
• Clients with less than $200,000 in assets who frequently change advisors often generate lower revenue and are more likely to leave.
Note that these explanations are generated by AI, taking into consideration many data points about your existing clients.
Attrition prediction, or detection, models are my favorite initial AI models, because they enable you to understand the quality of your data and build a strong foundation for other models. Another advantage of starting with attrition prediction is the ability to unite your entire organization around better understanding your clients, data and each other.
Once you understand and decide who your best clients are, you are ready to train a new AI model to find new leads that match these high-value clients’ attributes, and you stop wasting resources on wrong prospects.
Step 2: Collect The Right Data To Find Ideal Leads
The amount of data you have about your clients is a lot more than the data you have about your prospects. Therefore, to find prospects similar to your ideal client profile, you would use both internal and external data.
Internal data comes from your CRM and your wealth management platforms. It includes information such as the lead source, investment interest and experience, net worth, annual income and engagement score with your marketing outreach and content. You can also utilize external data sources from market intelligence providers to determine the prospect’s business ownership, real estate holdings, spending habits, trading activity and social media data.
When the cost of data acquisition is high, your ICP explanations can help you reduce such costs by focusing on the right set of clients you want to acquire.
Step 3: Create Your Explainable Lead Scoring AI Model
By combining these data sources together with the results of your acquisition efforts, you can train your lead scoring AI model to predict which leads are more likely to convert.
Explainable AI utilizing a decision tree surrogate model could include insights on how, or why, a lead would convert, for example:
• Internal Signals: rapid asset accumulation (exited company with $12 million payout), deep investment expertise ($3 million in brokerage accounts) and search for advisory services (attended two exclusive webinars, multiple blog post interactions).
• External Signals: Commented on a post about family office structures.
In addition to increasing trust in AI decisions, these explanations could lead to informed outreach actions, for example:
• Outreach by a senior advisor in wealth structuring for entrepreneurs.
• Tailored proposal for asset diversification.
• Private client dinner with other entrepreneurs.
Explainable AI, utilizing a linear model as a surrogate model, could show how much each factor, such as net worth, referral status, contributes to the lead conversion probability.
Then, based on the scores of the lead scoring AI model, you can eliminate the low-probability leads and focus your efforts on the high-probability leads in your ICP.
Client acquisition utilizing explainable AI, together with internal and external data sources, enables wealth management firms to understand and focus on their ICP, understand why each lead is valuable and also create highly personalized campaigns and outreach.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
link