Full Episodes | Insurtech Leadership Podcast

Introduction

Most carriers have stopped asking whether AI works. So why does so little of it ever leave the pilot?

Kurt Diederich has spent 25 years building the core policy, billing, and claims systems that property and casualty insurers run on. Two years ago he looked at AI and, in his words, thought it was a pile of trash. He kept retrying it every six months anyway, and the thing that flipped him was a bread recipe. He asked a model to cut it in half, it correctly turned tablespoons into one and a half teaspoons, and he went in with both feet.

That arc is the episode. Kurt walks through the adoption curve he sees in most organizations, from skepticism, to being wowed, to badly overestimating what the technology can do, and then the hard landing into execution where the roadblocks show up. His argument is that the software was never the constraint. The constraint is whether a company will restructure how it works, put the right people in front of the change, and actually measure what it gets.

Guest Bio

Kurt Diederich is the Chief Executive Officer and co-founder of Finys, a Troy, Michigan software company whose Suite covers policy administration, billing, claims, portals, and business intelligence for P&C carriers, mutuals, and state FAIR Plans. He started writing assembly-language games as a teenager on TRS-80s and Apple IIs, worked his way through college as a consultant, and has been an entrepreneur ever since. He spent the 1990s building custom claims and policy systems for carriers, then founded Finys in 2001 and went all in on insurance, building the platform on Microsoft .NET. Finys marked its 25th anniversary in 2026 and employs roughly 175 people.

Key Topics

-The four-stage adoption curve - Skepticism, then being wowed, then overestimating what AI can do, then the drop into execution where the real roadblocks appear.

-30% versus three to five times - Let people work the way they always have and you get a modest lift. Restructure the work into small pods and the number changes completely.

-The SDLC becomes the AI PLC - Kurt argues the software development lifecycle is turning into a product lifecycle, where the idea of developers sitting down to write code starts to go away.

-The 40-point measurement gap - Studies found developers predicting a 20% productivity gain while actually running a 20% deficit, and most organizations never measure at all.

-Champions rather than training - Training wears off because people default back to old habits. Kurt uses people who have already done it to coach teams through real work, repeatedly.

-Why carriers have it harder - Governance and regulatory exposure slow carrier adoption compared with a software company, and improper usage draws regulators.

-Spending AI dollars badly - Everyone reaches for the frontier model. Kurt watches an internal credits leaderboard, argues most problems are solved by cheaper models, and says AI should be used to build deterministic solutions rather than being run over and over.

Notable Quotes

"Developers were predicting they're getting a 20% productivity lift with AI when in reality they were at a 20% productivity deficit. So you've got a 40% gap between reality and perceived value."

"If you allow the developers to just work as they've always worked, what we find is you get maybe a 30% productivity lift. By restructuring the way you work, you see these huge productivity lifts."

"I don't think the typical worker is going to have that aha moment where they realize what they've been doing for the last ten, fifteen, or twenty years is wrong. They've been successful at it."

"The AI industry has kind of done themselves some harm by saying you're not going to need all the people to do the work anymore."

Resources

Guest:

Finys: https://finys.com/

Finys on LinkedIn: https://www.linkedin.com/company/finys/

Kurt Diederich on LinkedIn: https://www.linkedin.com/in/kurt-diederich/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

How much of your agency's new business comes from a source you don't control? Jake Cash spent five and a half years inside GEICO's national agency marketing operation, managing more than $500 million in ad spend, and watched the fastest-growing agencies ride a single channel up and then back down. On this episode of the Insurtech Leadership Podcast, he and host Joshua R. Hollander get into what it takes for a P&C agency to own its growth engine instead of renting it, and why the answer usually starts with the book of business the agency already has.

Guest Bio

Jake Cash is the Founder and CEO of Senryx Group, a Carmel, Indiana firm that builds marketing systems P&C agencies own and run themselves. At GEICO he led agency marketing strategy for more than 300 agencies nationwide, managed over $500 million in ad spend, and negotiated partnerships with Meta, Google, and NBA and MLB teams. He left that Fortune 500 seat (the same day his wife returned from maternity leave and his daughter started daycare) to build Senryx around a diagnostic-first model with no retainer lock-in: assess and plan, build the systems, then train the agency's own team to run them. His goal for most clients is self-sufficiency within six to twelve months.

Key Topics

-The five buckets of dependence - Carrier, channel, tactic, people, and vendor dependency, and how Senryx diagnoses which ones are capping an agency's growth.

-The 50% red flag - Why more than half of new business coming from one source is a vulnerability, and the first question every owner should be able to answer.

-From 100 to 800 policies a month - The 18-month rebuild of a Texas personal lines agency, and the channel mix (Google ads, Meta ads, cross-sell email, organic social) behind it.

-The book you already have - Retention runs seven to nine times cheaper than acquisition, and cross-sell revenue is already sitting in the existing client base.

-Owning your data - Lead vendors sell the same lead to five or ten agencies and share nothing, while owned ad accounts and email put the numbers in the agency's hands.

-AI: admin speed over content volume - Where AI actually helps an agency today, why unedited AI content fails, and Senryx's early tests running ads inside ChatGPT.

-Hiring for growth mindset - What roughly 30 hires in 10 years taught Jake, and why owners have to stop wearing every hat around the $1M revenue mark.

Notable Quotes

"If you don't know where you're getting new business from, then you're just flying blind."

"A lot of times there's kind of a secret pile of cash that's already sitting in your book of business."

"You're just getting the average of the internet as the reply."

"If we get to a point where they don't need us anymore, then we've done our job."

Resources

Guest:

Senryx Group: https://www.senryx.com/

Free two-minute Marketing Independence assessment: https://www.senryx.com/podcast

Jake Cash on LinkedIn: https://www.linkedin.com/in/jakecash/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

What happens when the personal lines book that carried an agency for years stops feeling safe? Rising home and auto rates, constant remarketing, and carriers pulling back have pushed a lot of independent agencies to look at commercial lines as the way to build a more resilient book. The appetite for the business has always been there. The submission process is what gets in the way.

On this episode of the Insurtech Leadership Podcast, host Joshua Hollander sits down with Michael Rabinowitz, who leads commercial lines product for EZLynx at Applied Systems. Rabinowitz started on the agency side when commercial rating meant loading a floppy disk of carrier rates every month, and he has spent nearly three decades trying to make the work simpler.

They get specific about why commercial has stayed hard, what changes when real-time carrier quoting moves inside the submission, and where AI actually helps an agent who is new to commercial.

Guest Bio

Michael Rabinowitz leads product management for the commercial lines business at EZLynx, part of Applied Systems, the largest provider of cloud software for insurance agencies in North America. His career in insurance spans nearly 30 years. He started on the agency side handling commercial submissions, then moved to the technology side around 2000 to work on some of the industry's early multi-carrier commercial rating. At EZLynx, he led the rearchitecture behind Instant Commercial Quoting, which brings real-time carrier rates into the EZLynx submission workflow. His focus has stayed consistent across all of it: build the tools he would have used as an agent, and make commercial simple enough that more agencies take it on.

Key Topics

-The resilient book - Why the personal lines hard market, with higher rates and heavier remarketing, is pushing agencies to diversify into commercial to stabilize the business.

-The four-times problem - How one small commercial submission can mean entering the same data into three or four carrier portals, often eating the entire commission on the account.

-Commercial by accommodation - Almost every personal lines agency already writes a few commercial policies for long-time customers, which makes commercial an extension of existing relationships rather than a cold start.

-Real-time appetite - Why carrier appetite now shifts day to day, and how live rating inside the submission changes which carriers get seen first.

-AI as an accelerant - How AI helps with class-code and NAICS lookups and catches misclassification, like the ice cream shop that is really a manufacturer, so newer commercial agents build confidence.

-Get out of the way - Rabinowitz's product philosophy that technology should cut context switching and re-keying, not add one more interface to learn.

-Carriers and friction - Why carriers that make real-time connectivity easy will earn the first look, and the ones that do not may get left off the submission.

Notable Quotes

"For me to be doing my job correctly required me to enter in the data four times. That is a tremendous amount of time wasted for maybe a two thousand dollar premium. I've probably eaten up my commission, if not more, just on that activity alone."

"I can't afford to have technology get in my way. Technology either helps me be more productive, or it causes me to lose opportunities."

"They're not an ice cream store, they're actually an ice cream manufacturer. If you misclassify a business and bind the policy and there's a claim, that's a headache down the road waiting to happen."

"If they have to go to five markets and only three are on real-time rating, they may go to those three first. As a carrier, you might get left off of a submission."

Resources

Guest:

EZLynx: https://www.ezlynx.com/

Applied Systems: https://www.appliedsystems.com/

Michael Rabinowitz on LinkedIn: https://www.linkedin.com/in/michael-rabinowitz-32998138/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

Why does an industry that talks about digital transformation still pay so many claims by paper check? The answer is not stubbornness; it is complexity, and it is worth billions. Josh Hollander sits down with Kevin Ostrander, Chief Revenue Officer at One Inc, to unpack what it actually takes to move carriers off checks, why multi-party claims payouts are the hard part nobody solved first, and what happens when enough carriers and payees are connected that the network, not the processing, is what carriers are buying.

Guest Bio

Kevin Ostrander is Chief Revenue Officer at One Inc, the insurance-specific payments network processing over $245 billion in annual payment volume for more than 300 carriers. He has spent nearly sixteen years selling enterprise technology into insurance, including six years at Thunderhead before its acquisition by Accel-KKR, and he joined One Inc ten years ago when the company had single-digit customers. He owns everything from new business acquisition through customer success.

Key Topics

-Why checks persist - More than half of claims payouts go to third-party vendors, lienholders, and mortgagees, and fifty states each add their own compliance rules, so single-party digital solutions never covered enough of the workflow.

-The malleable ROI case - Checks cost $6 to $25 each fully loaded, digital claim payouts land in minutes instead of days, and the same platform argument works whether the carrier's priority that year is cost, retention, or customer experience.

-Adjuster adoption decides outcomes - If issuing a digital payment adds manual steps, adjusters revert to checks, so One Inc pairs its technology with change management, FAQs, and training on the carrier side.

-One platform for pay-in and pay-out - Over 40% of One Inc's carriers now run both PremiumPay and ClaimsPay, giving policyholders one wallet across premium payments, refunds, and claims.

-The network is the moat - One Inc's escrow network covers 80%+ of escrowed homeowner premium and connects it to 300+ carriers, moving premium 11 to 12 days faster than the 12-to-15-day lockbox process.

-Disciplined expansion - Life insurance (AAA Life, Transamerica) and Canada (with Guidewire, which holds 80%+ of Canadian claims administration) came only after the P&C base was solid.

-Career advice from a first-time CRO - Kevin's path from single-digit customers to 300+ came down to experience you cannot skip, mentors, and recognizing product-market fit when you have it.

Notable Quotes

"Fifty plus percent of the payouts in the insurance industry actually go out to third-party participating vendors."

"The cost of a check fully loaded, including service, print, mail, et cetera, is anywhere from six dollars to twenty-five dollars a check."

"We're actually increasing the speed of premiums by sometimes eleven to twelve days."

"The one thing you can't skip in your career is experience. And you've got to find mentors that will help you through that process."

Resources

Guest:

One Inc: https://www.oneinc.com/

Kevin Ostrander on LinkedIn: https://www.linkedin.com/in/kevin-ostrander-6894ba2/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

Building the model is no longer the hard part of insurance AI. The hard part now is getting underwriters and carriers to actually adopt it. Josh Hollander sits down with Roger Ferrandis, Head of Partnerships at Sixfold, whose AI now runs inside carriers like Zurich, Guardian, and Skyward Specialty, to talk about how trust actually gets built: proof-of-concept by proof-of-concept, underwriter by underwriter, and market by market through partners who already have the relationships.

Guest Bio

Roger Ferrandis is Head of Partnerships at Sixfold, the AI underwriting platform. Originally from Spain, he spent seven years in the UK, where he co-founded WeAnalyze, an underwriting data startup he helped grow to 25 employees, $3 million raised, and clients on three continents before it pivoted to banking. He chose insurance over banking, moved to New York, and now runs Sixfold's global partner strategy across systems integrators, reinsurers, data providers, and cloud marketplaces.

Key Topics

-Adoption is the product problem now - Leadership teams know they need AI, hand it to someone to evaluate, and that person's reference point is ChatGPT making mistakes, so skepticism is the default starting position.

-How trust gets built in a POC - Sixfold runs known submissions through its AI, compares decisions with the underwriters, feeds their reasoning back into the model, and iterates until 90-95% of outcomes match, which is when confidence flips.

-Amplify, not replace - Sixfold is not chasing full automation; the goal is an underwriter handling four submissions in the time one used to take, with clients choosing their own comfort level on straight-through processing.

-The ROI stack - Clients measure 50% time saved per submission, quote-to-bind up 15%, and gross written premium per underwriter up 30%.

-Partnerships as market entry - Munich Re's Realytix platform, Microsoft's Azure marketplace, Adnovum in Switzerland, and Sollers Consulting in Poland each open doors that direct sales cannot, especially across Europe's twenty-plus selling cultures.

-Build versus buy, honestly argued - Sixfold's model has trained on nothing but underwriting for three years, and Roger's warning to carriers building in-house is that the technology shifts faster than an internal build can keep up with.

-Involve the underwriters - His advice to carrier leadership: do not delegate AI evaluation and hope for the best, because if underwriters do not buy in, nothing happens.

Notable Quotes

"We can sell to C-levels. But if underwriters are not happy with our solution, nothing's going to happen."

"With that iteration, we get to a point where at least ninety or ninety-five percent of the outcomes of the brain are the same outcomes the underwriter will have. When they see that level of accuracy, that's when we gain their confidence."

"If an underwriter used to deal with one submission every hour, we want them to be able to deal with four."

"By the time you build something yourself, the whole technology has changed so much that you need to rebuild it."

Resources

Guest:

Sixfold: https://www.sixfold.ai/

Roger Ferrandis on LinkedIn: https://www.linkedin.com/in/rogerferrandis/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

What happens when the lawyer who spent his nights cleaning up coverage disputes decides to build the software himself? Most insurtech is built by engineers learning insurance, and Dan Schuleman came at it the other way. He practiced insurance coverage law, watched a late claim letter turn into a bad-faith fight, and built Qumis so the people reading policies actually understand what the words mean.

Guest Bio

Dan Schuleman is co-founder and CEO of Qumis, the Chicago company building attorney-trained AI that reads and interprets insurance policies for brokers, underwriters, and claims teams. He spent his early career as an insurance coverage attorney at Am Law 200 firms, advising carriers and policyholders on high-stakes commercial claims, then became an early legal hire and Associate General Counsel at Kin Insurance, now a unicorn. Qumis raised an oversubscribed $4.3 million seed in February 2026, after a $2.2 million pre-seed, and its technology is used by five of the fifteen largest U.S. brokers, including NFP and Brown & Brown.

Key Topics

-Claims adjusters are practicing law. An adjuster without a law degree still reads a legal contract every day and forms an opinion on how a court would interpret it, which is the overlap Qumis is built around.

-The coverage letter that went out late. Dan traces Qumis back to a hotel roof claim that turned into a bad-faith dispute because the letter did not go out in time.

-The bench of digital experts. Ask Qumis a question and a lead agent assembles specialist agents that each examine the policy and then synthesize one cited answer, the way a well-resourced firm puts a team on a file.

-A 97% lawyer-agreement rate, with citations. Every output traces back to the source text and the reasoning behind it, so an adjuster can agree or disagree instead of trusting a black box.

-Where AI stops and a lawyer starts. Dan calls coverage interpretation one of the hardest things to automate, and he does not see humans leaving the process any time soon.

-Insurance's spreadsheet moment. He compares the shift to accountants and Excel, where the tabulating goes away and the judgment and creative work expand.

-Commodifying routine coverage counsel. Routine coverage questions that get outsourced to outside counsel are the part Dan expects AI to absorb first, changing when and how firms engage lawyers.

Notable Quotes

"AI isn't going to replace humans, but humans using AI will."

"The product is a promise, and then the promise is expressed in a whole bunch of legalese."

"I saw it play out in the claims context, where a comma could mean a million bucks."

"I have 500 pages of PDF on my desk, and I need to spend the next eight hours figuring out what the issues are and getting the letter out. We can turn that down into half an hour, with likely a more accurate output."

Resources

Guest:

Qumis: https://www.qumis.ai

Dan Schuleman on LinkedIn: https://www.linkedin.com/in/danielschuleman/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If this episode was useful, subscribe and leave a review. The Insurtech Leadership Podcast is on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

Captive premiums have crossed roughly $79 billion globally, so why are so many captive programs still managed in Excel and email? In this episode, Joshua Hollander talks with Illia Pinchuk, CEO and founder of DICEUS, about where manual workflows leak value in a captive program and what it takes to move a captive manager off spreadsheets. Listeners will hear how feature-based pricing changes the cost math for captives, why data onboarding is the real adoption barrier, and where AI actually helps today.

 

Guest Bio

Illia Pinchuk founded DICEUS in 2011 and has led the company for fourteen years, evolving from professional services for insurers, brokers, MGAs, and TPAs into ready-made products. DICEUS now offers 22 products across business lines, with captive insurance among its strongest segments, anchored by a captive management platform and a captive owner portal. Pinchuk trained as an engineer in mechatronics and robotics and has completed executive coursework through Harvard Business School Online.

 

Key Topics

-A young market with about 7,000 captives - Roughly half of the world's registered captives sit in the US, with Bermuda, Cayman, and a developing UK regime behind it, and business processes remain largely unstandardized.

-Why captive managers can't scale on spreadsheets - The consultant headcount a captive manager needs grows roughly in proportion to the captives it onboards, which turns a service business into a people business.

-Where value leaks in a manual captive - A/B fund premium allocation, loss runs arriving from 10 to 20 different TPAs in incompatible spreadsheets, investment reconciliation, and end-of-period compliance workbooks.

-Feature-based pricing, not module-based - Clients pay only for the specific features they use, which matters in a segment where the entire case for a captive is cost.

-The measurable payoff - Clients report the platform saves the work of one to one and a half business analysts or captive consultants, and policy issuance that once consumed a dedicated person compresses to about a week and a half.

-Key-person risk in Excel - When the one employee who knows the spreadsheets leaves, the knowledge leaves with them; built-in onboarding and AI guidance get a replacement productive in one to two weeks.

-AI needs clean data first - DICEUS runs the full extract, transform, load process itself, cleaning years of unstructured spreadsheet data before the platform's AI features can pay off.

 

Notable Quotes

"They cannot grow without adding new people. So it becomes more people business than just a service business."

 

"At the end of the month you're getting twenty completely different spreadsheets, and you somehow should make a magic work in order to calculate the loss ratio."

 

"Our system can save, for a traditional captive manager organization, at least one or one and a half consultants. But in reality, I think it's even more."

 

"They should pass through the digital onboarding step, just filling in all the details, and it should be automatically enrolled. Like, for example, if you want to open a bank account. That will be a big game changer."

 

Resources

Guest:

DICEUS: https://diceus.com/

Illia Pinchuk on LinkedIn: https://www.linkedin.com/in/illiapinchuk/

 

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

 

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

What happens to a $114 billion market when six people head toward retirement for every one person coming in behind them? That is the question hanging over the MGA channel, and it sits at the center of Vertafore's 2026 MGA Workforce and Technology Report. Josh Hollander sits down with Emily McGinn, who runs Vertafore's MGA & Wholesale business unit, to dig into what a survey of nearly 200 MGA leaders says about succession, operational discipline, and where AI is already paying for itself.

Guest Bio

Emily McGinn is SVP & General Manager of the MGA & Wholesale business unit at Vertafore, one of the largest insurance technology providers in North America, where she runs product development, professional services, and customer service for the segment. She came to insurance nine months ago from telecom and technology operations leadership at Zayo, Lumen, and Windstream, which gives her an outsider's read on an industry that keeps its people for decades.

Key Topics

-The demographic cliff - 67% of the MGA workforce is 44 or older, only 4% is under 28, and the ratio of workers 55-plus to workers under 25 runs about six to one.

-What walks out the door first - Carrier and agency relationships built over three or four decades, plus underwriting judgment that has seen full market cycles, are the things technology cannot replace.

-Growth is moderating, and it changes the playbook - MGA premium growth has cooled from double digits to single digits, so premium volume alone no longer covers for thin operational discipline.

-Operational excellence is now a carrier mandate - Carriers are demanding clean data, real governance, and bordereau reporting that ties out, and they are choosing MGA partners accordingly.

-Where AI pays for itself today - Vertafore's email ingestion agent reads unstructured submissions and fills the application automatically, turning 20 minutes of data entry into a three-minute review.

-Tech-first MGAs are table stakes, not a threat - The startups are pushing everyone to move faster, but the winners still need the industry foundation first and the technology on top.

2026 as the adoption inflection - 21% of surveyed MGAs use AI today and 50% plan to adopt soon, and Emily believes those numbers were already stale within months of the survey.

Notable Quotes

"The ratio is about six people 55-plus to one person under 25."

"It used to take you twenty minutes to read these PDFs and data-entry this all in. Now it takes seconds. You review it in three minutes to make sure it's right and move on."

"If you just start with the technology, you're probably not going to win."

"The MGAs that didn't make it, it was often a data issue. They lost the carrier's trust."

Resources

Guest:

Vertafore: https://www.vertafore.com/

Emily McGinn on LinkedIn: https://www.linkedin.com/in/emily-dempsey-mcginn/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

What makes an underwriter act on an AI's answer? Not the accuracy of the model, according to Stan Smith, but whether the person can see how the model got there. Josh Hollander sits down with the founder and CEO of Gradient AI, whose platform is trained on a contributory data lake of tens of millions of policies and claims, to talk about explainability as the last mile of AI adoption, what regulators actually want to see, and the A/B test that put a hard dollar figure on AI-managed claims.

Guest Bio

Stan Smith is the founder and CEO of Gradient AI, which builds AI that insurers use to underwrite risk and manage claims across both P&C and health. He started the business inside Milliman, bought it out in 2018, and has since raised roughly $90 million in growth capital and grown from about a dozen clients to several hundred. Before Gradient, he built a machine-learning startup that predicted supplier performance from pooled supply chain data, the same contributory model that now powers Gradient's data flywheel.

Key Topics

-Explainability is the last mile - A correct number the underwriter cannot interrogate gets ignored, and a number with visible reasoning gets used, disagreed with productively, and trusted over time.

-What regulators actually want - They are not auditing the math; they regulate inputs and outputs, with the sharpest focus on personal lines, like Massachusetts barring personal credit in personal auto underwriting.

-GLMs versus AI - Linear models stay popular because they are explainable, but they miss subtleties in the signal, and Stan argues that trade-off costs accuracy the industry does not have to give up.

-The $7 million A/B test - A large self-insured employer held part of a roughly 20,000-claim book out of Gradient's claims management and measured $7 million in savings on the AI-managed side.

-The contributory flywheel - Clients share data because Milliman-era trust made it safe, and the pooled data makes every client's models better, which is what in-house builds cannot replicate.

-The MVP trap - Carriers that build internally usually stall at a minimally viable product, over budget and behind schedule, while vendors iterating across hundreds of clients keep compounding.

-What got us here won't get us there - Stan's scaling mantra: priorities, execution discipline, and accountability have to change every year, without losing startup speed.

Notable Quotes

"If the person is not confident as to how the model came to that conclusion, they can just pass, even though the model might have given them some important directional information."

"They measured a seven million dollar improvement in their loss costs on the claims we were managing versus the claims in their A test."

"What they build in-house tends to be a minimally viable product. They've told me this. I haven't said it to them."

"My constant mantra to the team is: what got us here won't get us there."

Stan Smith

Resources

Guest:

Gradient AI: https://www.gradientai.com/

Stan Smith on LinkedIn: https://www.linkedin.com/in/stan-smith-5029246/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

Introduction

Most insurance executives have heard the AI pitch a hundred times by now, and most of those pitches collapse the moment they hit a real call center, a real compliance team, or a real broker's quoting workflow. The interesting question is no longer whether AI works in insurance. The question is what deployment method survives a regulated, audited, multi-channel customer operation, and which vendors actually know how a policy gets quoted, bound, and serviced.

In this episode, Joshua R. Hollander sits down with Pablo Molina, CEO and co-founder of Infinite Watch, to unpack what AI-native infrastructure looks like inside a carrier or broker. Molina argues for a method that starts with observability rather than automation, and for a vendor selection criterion that has shifted from technology to domain expertise and traceability.

Listeners will come away with a clearer view of how to sequence an AI deployment in a regulated business, why a hundred percent visibility into customer interactions is now table stakes, and what Molina sees coming next for the workforce.

Guest Bio

Pablo Molina is CEO and co-founder of Infinite Watch, an AI-native platform built for insurance and other regulated verticals. Infinite Watch operates three families of agents: conversational voice, data and document handling, and real-time business intelligence. The company came out of stealth in late 2025 and is running production deployments in personal lines, commercial lines, and payment collections across the US and Europe.

Before Infinite Watch, Molina was the founding CTO at CoverWallet, where he scaled engineering past 300 people and helped grow the business past one billion dollars in premiums before Aon acquired it in 2019. He is a software engineer with more than a decade of insurance domain experience.

Key Topics

• Observability before automation - Why Infinite Watch deploys insight agents first, ingesting calls, messages, and web interactions before any task gets automated. Operators need a true picture of customer interactions before deciding which parts to hand to an agent.

• From 1-5% audit samples to 100% visibility - Carrier and broker QA teams sample one to five percent of calls. Molina explains why semantic LLM coverage of every interaction changes the operating model for playbook adherence, cancellation analysis, and missed upsells.

• Where the moat actually lives in AI for insurance - With frontier models commoditizing fast, the differentiator is no longer the technology stack. It is domain expertise and built-in traceability that satisfies regulators on day one.

• Personal lines today, commercial lines next - How Infinite Watch is moving from automated payment collections and end-to-end claims into commercial-lines quote-and-bind, compressing days into minutes.

• AI replacement, not just augmentation - Molina pushes past the safe answer. He argues software engineering was the first profession to feel AI replacement because code is the easiest output to validate, and every profession follows.

• Founding team dynamics on round two - Putting a CoverWallet founding team back together with different titles, hiring small and senior, and building an AI-native engineering culture from day one.

• Three years out for Infinite Watch - Whether agents replace or complement legacy core systems, and which incumbents survive the transition.

Notable Quotes

"We first of all deploy AI insights agents that will ingest every customer interaction. We sit in the front office. Then we have a view of what the operations and customer interaction look like."

"There is no technology moat as such, like it used to be. The differentiators are not in the technology anymore, because people can catch up really fast."

"Audits and traceability is a built-in. There is no other choice. It is not a negotiation point."

"Professions are going to be replaced and reinvented entirely, like a hundred percent."

Resources

Guest:

Infinite Watch: https://infinitewatch.ai

Pablo Molina on LinkedIn: https://www.linkedin.com/in/pablomolinacandel/

Host & Organization:

Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/

Horton International (USA): https://www.horton-usa.com/

Insurtech Leadership Podcast: https://www.linkedin.com/showcase/insurtech-leadership-show

Subscribe & Review

If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.