How AI Is Transforming Long-Term Care Planning: A Conversation with Waterlily AI's Lily Vittayarukskul

TL;DR: Waterlily AI uses machine learning trained on 500+ million data points to compress LTC planning from weeks into minutes. The platform predicts personalized care journeys, matches policies in under one second, and shows advisors exactly where traditional illustrations fall short. For financial advisors, this means moving from generic cost estimates to precision planning that builds client trust through transparency.

A 45-minute LTC intake compressed to three minutes. Over a million policy comparisons in one second. Machine learning predicting a client's care journey down to phase-by-phase costs, ADL triggers, policy shortfalls.

This isn't theory. This is Waterlily AI.

Lily Vittayarukskul brought her NASA engineering background to a deeply personal problem. After watching her family navigate the chaos of long-term care during her aunt's cancer diagnosis, she built a machine learning platform trained on over 500 million data points. What she created is reshaping how advisors approach one of retirement's least-understood risks.

I sat down with Lily to understand the AI architecture behind Waterlily, how predictive algorithms replace guesswork, and why advisors who adopt this technology aren't working faster. They're working at a different level of precision.

What Is Waterlily AI and How Does It Work?

Walk me through the technology. You're using machine learning trained on over 500 million data points. What does this mean for an advisor sitting across from a pre-retiree?

Most advisors pull generic cost estimates, adjust for geography or care setting, then layer in insurance illustrations. The whole process runs on averages and assumptions.

Waterlily works differently. We trained our algorithms on claims data, actuarial tables, regional care costs, policy performance records, family caregiver patterns, health trajectories, ADL progression timelines. Over 500 million data points.

When an advisor inputs a client's age, health profile, family situation, financial picture, our machine learning models generate a personalized probabilistic care journey. Not an average. A prediction specific to this individual.

So instead of saying "the average 70-year-old needs X years of care," you're predicting this person's trajectory?

Exactly. The AI models when care will most likely begin. What type of care setting they'll statistically need first (home care, assisted living, memory care). How long each phase lasts. When they trigger ADL requirements for insurance benefits. What cost inflation looks like in their geography.

We're running probabilistic models across thousands of potential care paths, identifying the most statistically likely outcomes based on similar profiles in our dataset.

So this is fundamentally different from assuming 3 years of care at $100K per year?

Completely different. The precision gets better when we layer in the policy matching engine.

The Bottom Line: Waterlily's AI doesn't give you industry averages. It generates personalized care predictions by pattern-matching against 500+ million real data points.

How Does the One-Second Policy Matching Engine Work?

You mentioned the policy matching engine. This is where advisors lean in. How does this work?

This is where Waterlily becomes a competitive tool.

Traditional LTC quoting is manual, slow, limited by human capacity. An advisor pulls illustrations from 3-5 carriers, tests a few benefit configurations, presents what feels reasonable. This process is not instantaneous, and the back and forth on illustration questions and clarifications takes days or weeks over multiple meetings. You're also working from a tiny sample of what's available.

Waterlily's matching engine works differently. Once we have the client's predicted care journey, our algorithms evaluate over a million policy combinations in under one second.

A million combinations in one second. Break down what this means.

We test policy structures across dozens of carriers at once. Different daily benefit amounts. Benefit periods. Elimination periods. Inflation riders. Shared care options. Hybrid structures. We model how each configuration performs against this client's predicted care timeline, health profile, budget.

Here's what makes this intelligent: we're not showing all options. The AI identifies mathematically optimal matches. The policy structures providing best coverage for the specific care phases this client will most likely need, at price points they afford, from carriers most likely to approve them based on health profile.

You're layering in underwriting probability at the quoting stage?

Exactly. Most advisors don't have access to this. We trained our models on underwriting outcomes. We flag early: "This health profile creates issues with Carrier A, but Carrier B has flexible underwriting for this condition." Or "This medication triggers questions. Here's how to position it."

Advisors stop wasting time on quotes that look good on paper but won't survive underwriting. You're presenting realistic options from day one.

What This Means: Instead of manually comparing 3-5 carriers over days, Waterlily evaluates a million policy combinations in one second, flagging underwriting issues before they kill deals, confidence, and client trust.

What Happens During the Three-Minute Intake?

You've compressed a 30-45 minute intake to three minutes. What's happening from a technology standpoint?

From the advisor's view, it's simple. Answer a handful of questions about the client's age, health, family structure, financial picture, care preferences. Behind the scenes, machine learning does the heavy lifting.

The algorithms cross-reference inputs against our 500+ million data point training set. They identify similar profiles, predict care probability curves, model likely care settings and durations, estimate phase-by-phase costs adjusted for regional inflation. Then generate a Personal Care Blueprint mapping the most statistically likely care journey for this person.

Instead of 40 intake questions, the AI extrapolates from fewer data points?

Correct. This only works because of our training data depth. We're not guessing. We're pattern-matching against hundreds of thousands of real care journeys with similar characteristics, so not only do we only ask relevant questions based on what you tell us, but only ask about the most predictive data points, not average, irrelevant data points.

The output is a visual, phase-by-phase prediction of when care begins, what type, how long each phase lasts, what it costs. This becomes the foundation for policy matching, gap analysis, family conversation.

Key Insight: Three-minute intakes work because AI pattern-matches against massive datasets, not because you're skipping important questions.

How Does the Modeled Claim ROI Simulator Work?

Let's talk about the Modeled Claim ROI Simulator. What problem does this solve?

This makes policy performance transparent. Traditional illustrations show what a policy pays if everything activates perfectly. The Modeled Claim ROI Simulator shows what the policy will statistically pay based on this client's predicted care journey.

We model dollar-in, dollar-out efficiency. Premiums paid over time. When care begins. When the client triggers ADL requirements for benefits. Elimination periods. Daily or monthly benefit caps. Inflation adjustments. Phase-by-phase payouts. Remaining shortfalls. Death benefit impact if benefits aren't exhausted.

You're stress-testing the policy against the client's most likely care scenario?

Exactly. This reveals gaps traditional illustrations hide. Example: a policy with a $400,000 benefit pool might project only $260,059 in LTC payouts when tested against the client's predicted journey. Why? Waiting periods. ADL triggers. Benefit caps. Claim timing.

This isn't a knock on the policy. It's reality. Advisors who show this reality build more trust than those painting scenarios that don't survive real care needs.

Why This Matters: Traditional illustrations assume perfect benefit usage. Waterlily shows how policies perform against real care patterns, revealing gaps before families face them.

From Personal Experience to Data Science: How Waterlily Was Built

You came to this from watching your family navigate care during your aunt's cancer diagnosis. How did personal experience translate into an AI platform?

Personal experience showed me the problem. Data science solves it at scale.

Watching my family through that two-and-a-half-year care journey, I saw how unprepared we were. Not financially irresponsible. We were operating without a playbook. Who coordinates appointments? Who understands insurance? Who takes time off work? What's affordable? What happens when needs escalate?

This wasn't unique to us. Millions of families face the same chaos because LTC planning has been reactive, generic, disconnected from the shape of care people need.

How did you go from personal crisis to building machine learning models?

I brought my NASA engineering background to a human problem. Started researching. What does the data say about how care unfolds? When do people need help? What types of care do they use first? How long do episodes last? What causes families to break down financially or emotionally?

The data existed in claims records, actuarial tables, caregiver studies, health trajectories, regional cost databases. But nobody had synthesized it into a predictive tool advisors use at planning time.

We aggregated 500+ million data points, trained machine learning algorithms to identify patterns, built a platform generating personalized care predictions in real time.

The Origin Story: Waterlily began with a personal family crisis and an engineer asking, "What does the data actually say?"

Why Speed Matters in LTC Planning

You keep mentioning "three minutes" and "one second." Why does speed matter so much?

Friction kills planning momentum.

If intake takes 45 minutes, the advisor's exhausted before recommendations. If pulling quotes takes days or weeks, the client procrastinates, avoids, forgets why this mattered. If illustrations confuse, the conversation stalls.

Waterlily removes friction. Three-minute intake. One-second policy matching across a million combinations. Instant Personal Care Blueprint and Modeled Claim ROI Simulator. Digital applications auto-filling and e-signing.

You're compressing weeks into minutes?

Without sacrificing accuracy. We're increasing accuracy because AI evaluates far more variables and scenarios than humans process manually.

Speed creates a different client experience. Instead of "Let me get back to you in two weeks," the advisor says, "Based on what you told me, here's your predicted care journey, here are three policy structures fitting best, here's how each performs against your scenario, here's where gaps are."

That's the difference between conversations that fizzle and planning decisions that happen.

Speed Advantage: Removing friction keeps momentum alive. Faster isn't worse. It's more accurate because AI processes variables humans miss.

What's the Competitive Advantage for Advisors?

What's the tangible advantage of adopting Waterlily versus traditional LTC planning?

You move from reactive to predictive. Generic to hyper-personalized. Weeks to minutes. Guessing to mathematical modeling.

Practically:

• Have LTC conversations earlier because the process isn't burdensome
• Close more business because clients understand what they're buying
• Differentiate from competitors still manually pulling 3-5 carrier quotes
• Build deeper trust by showing math, not selling promises
• Uncover broader planning opportunities through transparent gap analysis leading to care reserves, home equity strategies, family coordination, estate planning

What about operations?

You reduce time on LTC cases without reducing quality. Avoid wasting time on policies that won't survive underwriting. Deliver client experiences that feel a decade ahead of competitors.

This is what happens when you replace manual processes with AI evaluating a million scenarios in the time one human pulls one illustration.

Advisor Takeaway: Waterlily gives you predictive precision, speed, and transparency that traditional methods physically cannot match.

What's Wrong with Traditional LTC Illustrations?

Most advisors use LTC illustrations. What's broken in those 20-40 page documents?

Traditional illustrations are technically useful but fail at translation. They're built around the policy, not the person.

A 20-40 page illustration shows premiums, benefit pools, inflation assumptions, death benefits, surrender values, internal rates of return. But clients still ask, "What does this mean for me if I need care?"

The biggest issue: many illustrations assume clean, idealized benefit usage. Client becomes claim-eligible, uses benefits efficiently, exhausts most or all of the policy. This makes IRR or value stories look more optimistic than real life.

What happens in reality?

Benefits activate unevenly.

Clients need help with one ADL before qualifying for claims. They have waiting periods. Start care at home, move into higher-intensity care later. Hit daily or monthly benefit caps. Leave part of the pool unused. Have large shortfalls even when the policy technically works.

Waterlily models this messiness?

Exactly. We don't pull illustrations. We model the client's most likely care journey, then test how the policy performs against that predicted journey.

How Dollar-In, Dollar-Out Transparency Works

In client examples, Waterlily models premiums paid, expected early/moderate/full-care phases, ADL eligibility, waiting periods, daily benefit caps, inflation, remaining benefit pool, phase-by-phase payouts, shortfalls, death benefit impact.

We might show projected policy coverage of $260,059 against much larger predicted care costs. Not pretending the policy value equals the illustrated benefit pool.

That's the translation layer advisors have been missing. Dollar-in, dollar-out policy efficiency based on personalized predicted LTC plans.

What about quoting and applying?

Waterlily's unique advantage: scale and speed.

Instead of manually pulling a handful of illustrations, Waterlily evaluates over a million quote combinations, layers automatic underwriting signals, identifies policy structures most likely fitting client health profiles, affordability, predicted care needs.

What takes years of manual comparison happens in a split second.

Advisors move from "Here's a policy illustration" to "Here's how this policy will perform for your family, given your likely care journey, claim timing, benefit activation, remaining out-of-pocket exposure."

Fundamentally different conversation.

Illustration Problem Solved: Traditional illustrations assume ideal benefit usage. Waterlily tests policies against real care patterns, showing where gaps appear before claims happen.

Does Transparency Kill Deals or Build Trust?

When you show clients their $260,000 policy benefit might still leave them exposed, how do they react? Does transparency kill deals or build trust?

Transparency almost always creates more trust, especially with serious clients. It kills the wrong deals. Strengthens the right ones.

If clients believe buying a $260,000 benefit means "care fully covered," that's not informed planning.

In practice, benefits pay differently depending on claim eligibility timing, whether they need one ADL or two, whether care starts at home, whether they hit daily benefit limits, whether care costs inflated faster than policy benefits.

In sample Waterlily models, policies projecting $260,059 total LTC payout still show major phase-by-phase shortfalls because eligible care costs exceed what daily benefits and remaining pools cover.

Clients seeing this don't say, "Then the policy's worthless." They say, "I finally understand what this does and doesn't solve."

That's a better foundation for financial decisions?

Advisors say, "This policy won't eliminate every dollar of risk, but it meaningfully reduces what your spouse covers, what your children absorb, what assets you'd otherwise liquidate."

This moves conversations from exaggerated certainty to real planning.

That's the difference between sales processes and fiduciary planning. Traditional illustrations unintentionally make products look cleaner than real life. Waterlily makes messiness visible. Not to scare clients. To motivate better planning decisions.

Clients trust advisors more when they show gaps, not only benefits.

Paradoxically, this makes insurance conversations stronger. Clients aren't buying based on rosy illustrations. They're buying because they understand exactly where policies fit inside broader care reserve and family resilience plans.

Trust Builder: Showing gaps doesn't kill deals. It builds trust with serious clients who want real planning, not sales pitches.

What's the Business Case for Transparency?

What's the business case when you choose transparency over cleaner stories?

Transparency creates higher-quality demand.

It makes purely transactional sales harder because you're not giving clients the clean story that "this policy solves long-term care." But it makes planning conversations stronger because clients understand the specific problem products solve.

Clients aren't pushed into policies. They see, concretely, how much risk remains if they do nothing, how much risk transfers if they buy coverage, where they still need care reserves or family plans.

This creates more conviction, not less.

Yes, this closes more business. More importantly, it closes better business.

Better-fit policies. Better expectations. Fewer surprises. Clients more likely believing you're acting as planner, not salesperson.

This helps uncover needs traditional illustrations miss. Clients might not buy the largest policy but buy the policy most efficiently covering the care phase mattering most. Or realize insurance is one plan part, coming back for broader retirement, estate, income, family planning work.

So that's Waterlily's business case?

We're not making insurance stories prettier. We're making them more accurate, personal, actionable.

In a world where clients are skeptical of being sold to, advisors saying, "Here's what this policy does well, here's what it doesn't solve, here's how we build around the gap," earn a different level of trust.

Business Reality: Transparency closes better business with serious clients while filtering out transactional buyers who weren't good fits anyway.

How Does AI Rebuild Trust After Industry Credibility Issues?

Traditional LTC policies went through rate actions, premium increases. The industry has credibility wounds. How does mathematical modeling rebuild trust?

The industry lost trust because too many clients were shown a simple promise: "Pay this premium, get this benefit." Reality turned out more complicated.

Premiums changed. Underwriting tightened. Benefits didn't activate how families assumed. Many people realized too late they'd bought products without fully understanding conditions around them.

Rebuilding trust isn't about newer, cleaner sales stories. It's showing the math. Giving advisors and clients tools to self-fund with confidence.

It's not about insurance. It's about holistic financial education and planning.

That's where Waterlily is valuable.

We show, transparently, how policies are expected to perform against clients' predicted care journeys: when care begins, when they become claim-eligible, how ADL requirements affect payout, how waiting periods work, how daily or monthly caps limit coverage, how much benefit gets used, what shortfall remains.

This level of transparency changes trust dynamics.

You're not saying, "Trust me, this is a good policy." You're saying, "Here's the math. Here's what this policy does well. Here's what it doesn't solve. Here's the remaining gap. Here's how we decide whether that tradeoff is worth it."

How AI-Driven Underwriting Insights Work

AI-driven underwriting insights help the same way.

Instead of letting clients emotionally commit to plans then discovering late that health profiles create underwriting issues, Waterlily surfaces those hurdles earlier.

This makes the process feel less opaque, less like a black box. Helps you avoid wasting client time on products that won't fit.

The business case and trust case are the same. Clients don't need more optimism. They need more clarity.

Waterlily helps replace the old credibility problem (oversimplified illustrations, surprise outcomes) with planning processes that are more personalized, more mathematically honest, much easier for clients to trust.

Rebuilding Trust: Show the math, surface underwriting issues early, replace promises with predictions. That's how AI repairs industry credibility.

What Becomes Possible When Advisors Adopt This Approach?

Last question. If every advisor adopted Waterlily's approach tomorrow (three-minute intakes, mathematically modeled care journeys, transparent gap analysis), what becomes possible in retirement planning that isn't possible today?

Conversations become earlier, more personal, more actionable.

In minutes, you show families what their likely care journey looks like, what it costs, where insurance helps, where gaps remain, how to protect people who'd otherwise become the default care plan.

Families make better decisions before crises hit. You lead one of retirement's most important conversations with confidence instead of discomfort. LTC moves from a product sale at the plan's edge to a core part of protecting dignity, independence, spouse security, legacy, family resilience.

That's the pre- and post-Waterlily shift. From "Do you want to buy a policy?" to "Here's the future we're preparing for, here's the plan protecting your family, your life, your legacy."

The technology shifts the entire industry forward.

Right now, most LTC planning happens too late, relies on generic assumptions, leaves families confused about what they're buying. Advisors know it's important but avoid it because the process is clunky, time-consuming, hard to explain.

When you remove those barriers with AI, when you model personalized care journeys in three minutes, match policies in one second, show transparent gap analysis, surface underwriting issues before they derail processes, LTC planning moves from the retirement conversation's edge to the center.

Families get real clarity on what their care might look like, what it costs, where insurance helps, where it doesn't. Advisors become professionals saying, "Here's the math, here's the gap, here's how we build around it."

This isn't better planning. It's a different level of advisory service entirely.

And that's what Waterlily delivers?

Machine learning trained on 500 million data points. Algorithms running a million policy comparisons in under a second. Predictive models turning chaos into clarity. Advisors who finally lead LTC conversations with the same confidence they bring to Social Security or investment allocation.

Human-first, AI-powered.

That's the future of retirement planning.

Frequently Asked Questions About Waterlily AI

How does Waterlily AI predict care journeys more accurately than traditional methods? Waterlily uses machine learning trained on 500+ million data points (claims data, actuarial tables, regional care costs, policy performance records, family caregiver patterns, health trajectories, ADL progression timelines). Instead of generic averages, it generates personalized probabilistic care journeys by pattern-matching against hundreds of thousands of real care scenarios with similar characteristics.

What makes the one-second policy matching engine different from manual quoting? Traditional quoting involves manually pulling 3-5 carrier illustrations over days or weeks. Waterlily's AI evaluates over a million policy combinations in under one second, testing different daily benefit amounts, benefit periods, elimination periods, inflation riders, shared care options, and hybrid structures across dozens of carriers. It also layers in underwriting probability, flagging which carriers are most likely to approve based on the client's health profile.

How does the Modeled Claim ROI Simulator reveal gaps in coverage? Traditional illustrations assume ideal benefit usage (clean claim eligibility, efficient benefit exhaustion). The Modeled Claim ROI Simulator stress-tests policies against the client's predicted care journey, modeling premiums paid, ADL eligibility timing, waiting periods, daily benefit caps, inflation, phase-by-phase payouts, and remaining shortfalls. This shows the dollar-in, dollar-out efficiency based on how benefits will actually activate in real scenarios.

Does showing clients policy gaps hurt sales? Transparency builds trust with serious clients while filtering out transactional buyers. When clients see realistic projections (for example, a $400,000 benefit pool projecting $260,059 in actual payouts due to waiting periods, ADL triggers, and benefit caps), they don't say the policy is worthless. They say they finally understand what it does and doesn't solve. This moves conversations from sales pitches to fiduciary planning.

How does Waterlily address the LTC industry's credibility problems? The industry lost trust through rate actions, premium increases, and benefits not activating as families assumed. Waterlily rebuilds trust by showing the math transparently, surfacing underwriting issues early (before clients emotionally commit), and replacing oversimplified promises with realistic predictions. Advisors aren't saying "trust me." They're saying "here's the math, here's the gap, here's how we decide if the tradeoff is worth it."

What's the practical time savings for advisors using Waterlily? Traditional LTC planning involves 30-45 minute intakes, days or weeks pulling carrier quotes, manual comparisons of 3-5 carriers, and lengthy client education on confusing illustrations. Waterlily compresses this to a 3-minute intake, 1-second policy matching across a million combinations, instant Personal Care Blueprint generation, and digital applications that auto-fill and e-sign. What took weeks now takes minutes without sacrificing accuracy.

Who is Waterlily AI designed for? Financial advisors who want to move from reactive, generic LTC planning to predictive, hyper-personalized strategies. Particularly valuable for advisors working with pre-retirees (ages 55-65, people 5-10 years from retirement) who need to coordinate pensions, Social Security, and LTC planning into comprehensive retirement strategies. Also benefits advisors who've avoided LTC conversations because traditional processes were too time-consuming or confusing.

How does Waterlily integrate with existing advisory practices? Waterlily provides the Personal Care Blueprint (visual care journey prediction), Modeled Claim ROI Simulator (transparent policy performance testing), mathematically optimal policy matching, AI-driven underwriting insights, and streamlined digital applications. These tools don't replace advisors. They give advisors precision, speed, and transparency that manual processes physically cannot match, enabling earlier LTC conversations and uncovering broader planning opportunities (care reserves, home equity strategies, family coordination, estate planning).

Key Takeaways

• Waterlily AI transforms LTC planning from weeks to minutes using machine learning trained on 500+ million data points to predict personalized care journeys, not generic averages.
• The one-second policy matching engine evaluates over a million combinations, layering in underwriting probability to present realistic options advisors don't waste time on quotes that won't survive approval.
• The Modeled Claim ROI Simulator reveals gaps traditional illustrations hide by stress-testing policies against predicted care scenarios, showing dollar-in, dollar-out efficiency based on real benefit activation patterns.
• Speed removes friction that kills planning momentum. Three-minute intakes, instant policy matching, and digital applications keep conversations moving from exploration to decision.
• Transparency builds trust with serious clients. Showing realistic policy performance (including gaps) doesn't kill deals. It filters transactional buyers while strengthening relationships with families who want fiduciary planning.
• AI rebuilds industry credibility by replacing oversimplified promises with mathematical modeling, surfacing underwriting issues early, and giving advisors tools to show the math instead of selling the story.
• This isn't faster planning. It's more precise planning. AI evaluates variables and scenarios humans physically cannot process manually, moving advisors from guesswork to predictive precision.

About Lily Vittayarukskul

Waterlily's CEO and Co-Founder, was on the path to become an aerospace engineer at an exceptionally young age, a patented inventor, starting college at 14 and interning at NASA by 16. But her life took an unexpected turn when her Aunt, a cornerstone of her first-generation immigrant family, was diagnosed with terminal stage colon cancer. Post chemotherapy her family was shocked to learn that health insurance doesn't cover long-term care costs, driving their family to provide care from the home, stepping away from work and school to take on the physically demanding tasks of daily caregiving. This devastating event revealed the profound gaps in the healthcare preparedness of most families, leaving lasting impacts on her own. Determined to ensure a future where no one else would face the same hardships, Lily shifted her focus from aerospace engineering to healthcare innovation.Graduating from UC Berkeley with a degree in Genetics and Data Science, Lily led product and engineering teams at multiple startups before founding Waterlily, a company dedicated to making healthcare and its costs accessible, understandable, and empowering for all.

About the Author

Chris Hensley is a financial advisor, podcast host, and author of Digital Kaizen: Voice-First Thinking and AI Systems for Continuous Improvement. With more than two decades of experience guiding clients through complex retirement and financial planning decisions, Chris focuses on judgment, pattern recognition, and systems that scale without losing humanity.

Digital Kaizen launched in April 2026 and hit #1 New Release in Business Office Skills on Amazon. The book explores how to work, think, and create in an AI-enabled world — blending human wisdom, voice-first workflows, and small, honest loops of improvement.

This blog is part of an ongoing series of conversations with financial advisors and fintech leaders using AI in practice. Those conversations are feeding the research for the next book in the series: Digital Kaizen for Financial Advisors, coming later this year.

Get the Book

→ Ebook ($4.99): https://a.co/d/0ihi9Y9F

→ Paperback: https://a.co/d/0hHTKDCN

Listen to the Full Conversation

The podcast version of this interview drops [TBD] on the Money Matters -Digital Kaizen series podcast — available on Apple Podcasts, Spotify, and wherever you listen.

Want More?

Grab the free Digital Kaizen Starter Guide for tools, frameworks, and insights from the book: https://digital-kaizen-book.kit.com/6a16de43b8