The Building Isn’t $50 Million. It’s $500,000. The Data Just Came In Wrong.

This episode brought two guests together for the first time on the show: Wendy Aarons-Corman, CEO of OWIT Global, and Barry Bablin, SVP of Actuarial Services at Farm Family Insurance (a subsidiary of American National) — a genuine practitioner-and-vendor pairing built around one of insurance’s least glamorous but most expensive problems: bad data.

Wendy’s path into the vendor side was itself an accident of temperament. She started as a software developer writing underwriting systems for specialty carriers, and by her own admission never planned to end up on what she half-jokingly called “the dark side” — the vendor side she’d had limited respect for while sitting on the carrier side of the table. Joining Duck Creek Technologies as an unusually extroverted developer changed that trajectory: she moved into business development almost by accident, bringing a genuine practitioner’s understanding of the underlying technology problems with her.

Barry’s perspective comes from the opposite side of that same table. A 35-year actuary, he spent 28-29 years in American National’s standard lines division before moving roughly five years before this recording into the company’s specialty markets group — a book of business produced almost entirely through MGAs and direct relationships with businesses like finance companies and furniture retailers, rather than through internal policy administration systems. That shift exposed him directly to a class of problem he’d never had to deal with in standard lines: a building recorded as a $50 million exposure that was, on inspection, actually worth $500,000 — bad inbound data, not a real risk.

In Episode 66 of InsurTechTalk, Wendy and Barry covered why bordereaux data integrity has quietly become one of insurance’s most expensive unsolved operational problems, what it would actually take to onboard or offboard an MGA relationship in 90 days, and why the industry’s growing data exchange ecosystem means every carrier is now effectively inheriting its partners’ data problems on top of its own.

About the Guests

Wendy Aarons-Corman is CEO and co-founder of OWIT Global, founded roughly four years before this recording by several Duck Creek Technologies alumni alongside other insurance industry veterans. OWIT provides a no-code platform for ingesting, cleansing, and normalizing bordereaux data flowing between carriers, reinsurers, and MGAs, along with additional tools extending legacy core systems (portals, rating, document generation) — described internally as “filling the gaps.” Barry Bablin is SVP of Actuarial Services at Farm Family Insurance, a subsidiary of American National, where he leads actuarial work within the company’s specialty markets division after nearly three decades in its standard lines business.

Why Bordereaux Data Is a Uniquely Painful Problem

Wendy’s framing of OWIT’s core focus: bordereaux management — the regular exchange of policy, claims, and premium data between an MGA (or reinsurer, or carrier) and its distribution or risk partners — is a genuinely insurance-specific problem that generic ETL (extract, transform, load) tooling doesn’t solve well, because it requires real domain understanding of policy data, claims data, cash reconciliation, and contract management, not just generic file transformation.

Barry’s account from the carrier side made the cost of getting this wrong concrete: in standard lines, he’d never had to think seriously about data integrity, because policy data entered his own internal systems directly. In specialty markets, ingesting data from dozens of external MGA partners, he began routinely encountering basic data quality failures — numbers transposed with letters, a decimal point effectively in the wrong place, values that simply didn’t hold up to a sanity check. His blunt summary: if he had a nickel for every one of those errors, he’d be a wealthy man.

The 90-Day Onboarding Dream

Barry described a specific, long-standing internal ambition at American National’s specialty markets group: the ability to fully onboard — or offboard — an MGA relationship within 90 days. Years into pursuing it, that goal remains a genuine stretch, and the reason, in his telling, isn’t really about data volume — it’s the friction of establishing a working electronic connection with a new partner absent any shared set of data exchange norms. Every MGA relationship today carries its own bespoke negotiation over file formats and data layout, and that friction cuts both ways: it makes it costlier for a carrier to recruit a new MGA, and costlier for an MGA to shop a new carrier, even when the underlying business relationship would otherwise make sense.

Barry called Wendy’s pitch — building toward a standardized, plug-and-play data exchange layer — the “secret sauce” that could genuinely close that gap: an industry-recognized standard set of variables and layout for a given product (his example: a defined set of roughly 120 variables) that any carrier or MGA working in that product line builds to, removing the bespoke negotiation entirely and letting both sides treat onboarding a new partner as a largely mechanical, low-friction process.

Why ACORD Doesn’t Actually Solve This

Asked directly whether ACORD — the insurance industry’s existing data standard — already addresses this, both guests were clear it doesn’t, at least not in practice. ACORD functions more as a general reference standard (form layouts, XML structures) than an enforced, universally adopted norm for MGA-to-carrier bordereaux exchange specifically. In practice, each MGA tends to have its own established data template, carriers request their own preferred format, and when the two don’t reconcile cleanly, the practical resolution is often the carrier’s business team simply saying “just give me what you’ve got” — meaning carriers frequently receive incomplete data by default, not because anyone intended that outcome, but because getting the deal done takes priority over enforcing a clean data standard nobody is actually required to follow.

OWIT’s Approach: Meet the Data Where It Is

Rather than trying to force the industry to standardize its outbound formats — a genuinely unrealistic ask given how entrenched existing processes are — OWIT’s platform is built to ingest whatever file type a partner already produces (Excel, fixed-format files, with PDF ingestion in development) and handle the cleansing and normalization on OWIT’s side. Wendy’s reasoning: getting people to change deeply established habits is hard, so the practical path is supporting what already exists today, while leaving room for more standardization to emerge naturally over time as the underlying tooling matures.

The Real Cost Isn’t Just Bad Data. It’s the Loop Around Fixing It.

Wendy pushed back on the idea that the core cost of dirty bordereaux data is simply inaccurate numbers reaching a report or board deck. The bigger, less visible cost sits in the operational loop required to fix it: manual review, back-and-forth communication asking a partner to correct a specific error, re-submission, re-review — a cycle that routinely delays when data actually lands in a carrier’s system. Barry gave a concrete, lived example of the downstream effect: a data file that should have loaded in March, delayed by a correction cycle, doesn’t actually get processed until April — producing a misleading trend line where March premium looks artificially low and April looks artificially inflated, purely as an artifact of a data-quality delay rather than any real business change. Anyone reviewing that trend later, without knowing the underlying cause, risks drawing a genuinely incorrect conclusion about what actually happened in the business that month.

Wendy noted this dynamic is increasingly pushing prospective clients to be proactive rather than reactive — companies growing their book of business are recognizing they simply can’t scale their current manual data-fixing process indefinitely. She cited a large global reinsurer OWIT had recently signed specifically because its internal team could no longer keep pace with the sheer manual effort required to cleanse incoming data at its scale.

Data Problems Used to Be Internal. Now They’re Inherited.

Both guests connected this to a broader historical pattern: data integrity issues have always existed inside insurance companies, even purely internally, well before external data exchange with MGAs and reinsurers became so central to specialty and program business. What’s changed is scale and direction — as carriers ingest more data from more external partners (and, looking further ahead, as more automated data exchange infrastructure, potentially including blockchain-based approaches, becomes more common), a carrier effectively absorbs its partners’ data problems on top of its own pre-existing ones. Wendy’s pointed observation: the industry has historically focused almost all of its technology investment on policy administration, billing, and claims — data integrity as its own discipline has been comparatively neglected, even as it becomes a larger and more consequential problem.

Why the Actuarial Side Feels This Acutely

Barry connected the data integrity problem directly to his own pricing work: identifying genuine new pricing opportunities in a specialty line — “building a better mousetrap,” in his words — depends entirely on clean underlying data. His observation about where carriers actually invest rigor: essential financial-statement-driving data elements get real scrutiny and validation on a strict schedule, because producing accurate monthly financials is a hard organizational requirement. Other operational and business metrics — the ones actually useful for more sophisticated pricing analysis — routinely receive far less validation rigor, with many carriers lacking even basic sanity checks (is this a valid zip code for this state, is this value within a reasonable range) on the data flowing in. That gap between what’s rigorously checked and what actually matters for advanced pricing work is exactly where a solution like OWIT’s adds value, by surfacing and correcting those unchecked data quality issues before they reach analysis.

Rising Data Volume, From Every Direction

Beyond bordereaux specifically, both guests pointed to a broader trend: individuals now generate far more personal data than ever (phones, wearables, smart home devices, connected vehicles), specialty and exotic product lines increasingly require genuinely new categories of data that didn’t exist as inputs before, and a growing number of intermediary parties (TPAs, underwriting partners, and others) touch any given transaction — all compounding the volume and complexity of data any carrier has to manage well.

Defining “Insurtech”: Carrier or Vendor?

The conversation touched on a genuine definitional disagreement in how the industry uses the word “insurtech.” Wendy’s original mental model: an insurtech is a genuine insurance company — like Lemonade — built with modern, streamlined technology from the ground up, not a vendor selling into the space. Barry noted his own internal team tracks the term more loosely, applying it to a wide range of vendors, particularly on the claims side, building tools to make claims processing meaningfully more efficient (photo-based claims intake and similar innovations among the examples he’d encountered).

Positioning Against No-Code, Low-Code, and Data Enrichment Players

Asked how OWIT positions itself relative to adjacent categories — general-purpose low-code/no-code platforms, and data enrichment vendors — Wendy was candid that she considers “no-code/low-code” largely a generational buzzword rather than a fixed technical category, predicting the industry conversation would likely shift toward AI/ML framing within a year or two. Her more substantive point: OWIT’s platform is built on insurance-specific microservices — modular components purpose-built around real insurance data and processes, not generic drag-and-drop tooling repurposed for insurance. She drew a distinction between OWIT’s no-code approach and something like Duck Creek’s own configurator, which she considers more genuinely low-code given its underlying complexity, while noting that customers ultimately buy outcomes, not a specific technical label — a low-code tool that reliably does exactly what’s needed, ships quickly, and is priced right is a perfectly reasonable choice even when a “no-code” alternative exists.

Core System Ecosystems: The Sun and Its Orbiting Planets

Asked about core system vendors like Guidewire, Duck Creek, Majesco, and Sapiens building out partner ecosystems rather than trying to build every capability themselves, Wendy was unambiguous that this is the right strategic move: ripping and replacing a legacy core system is prohibitively expensive, so the smarter path for both incumbents and their surrounding vendor ecosystem is letting core platforms function as durable backend infrastructure while specialized vendors like OWIT extend and enrich what they can do. Her metaphor: the major core system platforms function like a sun, with an ecosystem of partner vendors orbiting around them — vendors that don’t build this kind of open ecosystem risk eventually being displaced or sunset entirely.

Barry’s complementary observation from the carrier side: the specialty and program space increasingly runs on a growing number of small, highly focused “micro-provider” TPAs and MGAs, each doing one narrow piece of the underwriting or claims process exceptionally well — which is exactly the plug-and-play integration capability this kind of ecosystem approach depends on.

Bridging Between Different “Solar Systems”

Extending the metaphor, I asked whether OWIT’s real value is bridging between different core system “solar systems” entirely — a genuinely common problem for larger carriers that, through mergers and acquisitions, end up running several different policy administration systems that were never designed to talk to each other. Wendy confirmed that’s a real and distinct value beyond simple integration: plenty of vendors can wire two systems together, but wiring them together doesn’t guarantee clean data flows between them. She cited a concrete existing OWIT client: an insurance company using OWIT’s bordereaux tooling specifically to pull data out of multiple disparate internal policy systems, cleanse and normalize it, and feed the result into a centralized operational data store for reporting — precisely because the carrier’s own internal source systems weren’t producing clean data to begin with.

Recommendations: Growth Mindset and an Ice Castle

Wendy’s closing recommendation was Carol Dweck’s research and writing on growth mindset — the Stanford psychologist’s work distinguishing a growth mindset (the willingness to learn, make mistakes, and improve) from a fixed mindset (avoiding a problem specifically out of fear of failing at it). Wendy connected it directly to a second master’s degree she completed in organizational behavior shortly before the pandemic, and argued the concept applies as much to companies and their leadership culture as to individual learners — and that she sees the broader insurance industry itself gradually shifting toward more of a growth mindset.

Barry, recording from upstate New York in the depths of a February winter, recommended a seasonal attraction: Ice Castles, a temporary structure built entirely from ice (with locations including Lake George, New York, as well as New Hampshire and Minnesota), featuring slides, benches, and mirrored rooms, dramatically lit at night — a genuinely impressive, museum-like winter activity for anyone in a cold climate looking for something worthwhile to do before spring arrives.

Key Takeaways

• Bordereaux data integrity is a genuinely insurance-specific problem — cleansing, normalizing, and reconciling policy, claims, and cash data across MGAs, carriers, and reinsurers — distinct from generic ETL tooling because it requires real domain knowledge of insurance data structures
• A long-standing industry ambition to onboard or offboard an MGA relationship within 90 days remains difficult to achieve mainly because of a lack of shared data exchange standards, not data volume — every new partner relationship currently requires its own bespoke format negotiation
• ACORD, despite being the industry’s reference data standard, doesn’t function as an enforced norm for bordereaux exchange in practice — individual MGA and carrier data templates routinely diverge, and carriers often accept incomplete data rather than delay a deal over formatting disputes
• The real cost of dirty bordereaux data isn’t just inaccuracy in a report — it’s the operational loop of manual review, correction requests, and resubmission, which can delay when data actually lands in a system and produce misleading trend lines that lead to genuinely wrong business conclusions
• As carriers exchange more data with more external partners (MGAs, TPAs, reinsurers), they increasingly inherit their partners’ data quality problems on top of their own — a scaling challenge that’s historically received far less technology investment than policy, billing, and claims systems
• Core system vendors (Guidewire, Duck Creek, Majesco, Sapiens) building open partner ecosystems rather than attempting to build every capability themselves is, in both guests’ assessment, the correct strategic response to how expensive and disruptive full legacy system replacement remains
• Genuine differentiation in the crowded low-code/no-code and data-enrichment vendor landscape comes from insurance-specific domain depth (purpose-built microservices around real insurance data) rather than generic drag-and-drop tooling adapted after the fact for insurance use cases