Data Isn’t Wrong. It’s Just Never Synchronized.

Kabir Syed insists he’s not an entrepreneur, and says so every time he introduces himself. He came from India in 1995 for an MBA, joined Johnson & Higgins — later absorbed into Marsh — and stayed 17 years, cycling through roles because, as he puts it, someone let him play with data even though he wasn’t a broker, an actuary, or a programmer. He credits that early trust as the most important thing that happened to him professionally.

He left Marsh in 2012 to start RiskMatch, built on a specific observation about how commercial insurance actually gets placed: a broker doesn’t underwrite, a carrier does, and the connection between the two runs on relationships and lunches rather than any structured process. RiskMatch’s premise was that roughly 80% of risks share enough common structure to be matched systematically, even if everyone in the industry insists their risk is unique. RiskMatch grew 200-250% a year and was acquired by Vertafore (a Bain Capital company) after three and a half years.

He tried retirement. COVID made that untenable — he didn’t like spending his days on school pickup and honey-do lists — so he started ennabl in January 2021.

In Episode 125 of InsurTechTalk, Kabir and I covered why broker data is never actually wrong, just never synchronized, why giving humans the wrong task is the industry’s real problem, and why he refuses to chase a billion-dollar outcome on principle.

About Kabir Syed

Kabir Syed is the CEO and co-founder of ennabl, a data platform for insurance agents and brokers that synchronizes information across the disconnected systems a brokerage typically runs — placement systems, CRMs, agency management platforms — so that data entered once by one person in one role reflects consistently everywhere else. He previously founded and sold RiskMatch after 17 years at Marsh (via Johnson & Higgins). ennabl serves 70 broker and agency clients ranging from $3 million to $4 billion in revenue, and does not sell directly to carriers.

The Actual Name Story

ennabl was meant to be “Enable” — as in, enabling brokers to work faster, cheaper, easier. The name was unavailable, already owned by a coupon company. Kabir’s solution was to lean into an Indian-accented spelling, betting correctly that voice assistants and search engines would still resolve it to “enable” when asked. The guiding principle came before the name, which he considers the right order of operations for naming a company.

Why Broker Data Is Never Wrong, Just Never Synchronized

This is the conceptual core of the business, and worth stating precisely because it reframes what looks like a data-quality problem as something else entirely.

• An account manager enters data into one system. A producer enters related data into a separate CRM. A third person enters data into a placement system. Each is accurate at the moment of entry, from that person’s specific vantage point
• Because the systems don’t talk to each other, none of it stays synchronized as things change — the result is partial truths scattered everywhere, not incorrect data
• ennabl’s job is to synthesize those partial truths into one current, consistent picture, rather than treating any single system as the source of truth

The second half of the fix is arguably more important: stop asking brokers to manually enter data that can simply be purchased. Company address, firmographic details, demographic descriptors — all available externally. Every manual entry point is a future correction. ennabl treats certain categories of data entry as something that should never have been asked of a human in the first place.

The third piece is what the clean data is actually for — Kabir was explicit that unused data is just a statistic. ennabl orients its clients’ data toward three outcomes: making more money from a carrier, suggesting more to a client, and having staff spend more time on interpretation and less on re-entry.

Small Brokers Beat Large Ones on Data Cleanliness

I expected the opposite answer and got corrected. In Kabir’s direct experience, smaller brokerages have cleaner data and more automation than large ones — for structural reasons, not talent.

• Large brokerages have historically grown by acquisition — buying agencies and absorbing their books without ever unifying the underlying data infrastructure. The top three to five US brokers spend hundreds of millions trying to retroactively fix this; everyone below that tier is only just starting, because it was previously too expensive to attempt
• Small brokerages don’t add headcount casually — every hire has to be justified, so processes stay simple and roles stay tightly defined, with less of the middle-management layer that introduces inconsistency
• The tradeoff is real: smaller shops may have less sophistication, but the constraint of staying lean keeps the data clean almost by necessity

The Actual Problem Is a Human One

Asked point-blank whether this is a data problem or a human problem, Kabir’s answer reframes the whole conversation: it’s giving humans the wrong thing to do.

His analogy is a car dealership that makes a salesperson fill out 85 forms before completing a sale. The salesperson does it once, hates it, and actively avoids doing it again — which means the underlying objective (accurate, current data) never actually gets served. Applied to commercial renewals: if shopping a policy to the market means completing a lengthy separate form for every carrier, most brokers will simply renew with the incumbent rather than repeat the process, even when a better market exists. The friction itself, not any strategic decision, determines the renewal.

A Sacred Cow Worth Revisiting: Carriers That Won’t Take You Back

Kabir raised a genuine industry quirk with no clean justification: some carriers, after non-renewing a book of business, won’t take it back even years later when conditions have changed — behavior that would be unthinkable from almost any other kind of vendor. His broader point isn’t that carriers are wrong to manage risk appetite; it’s that the industry lacks a trusted, consistent mechanism across brokers, carriers, and regulators for revisiting these decisions as conditions change, so the friction just becomes permanent by default.

Automating the Renewal Sequence, Not the Relationship

We spent real time on a distinction Kabir was careful to draw: this isn’t about disrupting the broker-carrier relationship or replacing underwriting judgment with an algorithm. It’s about resequencing work that is already effectively data-driven.

• When a broker brings a risk to a carrier, the carrier is already pricing it with data science regardless of the personal relationship — the relationship may earn marginal goodwill, not a fundamentally different price
• Commoditized products — not complex ones — could clear through something closer to a structured exchange rather than a manual, one-off submission process
• His analogy: public company earnings already arrive in a standardized financial statement that analysts interpret and adjust for context; insurance pricing on commoditized products could follow the same pattern — a structured baseline plus human interpretation layered on top, rather than a fully manual submission built from scratch every time

The point isn’t disruption for its own sake — it’s that the industry does the same underlying analytical work today, just later and more expensively than it needs to.

Starting to Digitize Claims

ennabl’s newest area is claims, and Kabir’s description of the status quo was blunt: loss run data arrives from carriers as PDFs, and an account manager’s actual job, in practice, includes retyping that PDF into Excel — a manual, error-prone step sitting inside a genuinely multi-billion-dollar TPA and claims-adjusting ecosystem.

Two reasons this matters beyond accuracy:

• Placement strategy — if a prospect (his example: a restaurant that also operates a delivery truck) is a structurally high-loss risk because of the truck rather than the restaurant, that should inform where the broker places the business — because contingent commissions, which can represent 5-10% of a smaller broker’s profit, are driven by new business volume, retention, and loss ratio. A broker controls the first two and should place risk deliberately with the third in mind
• Cross-broker learning — Kabir’s view is that the industry operates as disconnected, closed islands that could benefit from sharing which carriers are genuinely good homes for high-loss business, without that being competitively sensitive information

Advice: Learn to Love the Work, Don’t Chase the Passion

Asked for closing advice, Kabir gave an answer that inverted the standard “follow your passion” framing. His actual passion was travel writing — pursuing it would have left him with no money and little influence. His approach instead was to learn to love what he was doing, likening it to an arranged marriage: not chosen from infatuation, but grown into deliberately. His point is that commitment to the work produces the passion, not the reverse.

He extended that into a pointed critique of startup culture broadly — too much noise, too many decks pitching a $10 billion opportunity on day one. His own approach at ennabl and RiskMatch: solve a genuinely $10 million problem first, let it lead naturally to the next connected problem, and be honest with investors that the company was never built to become a billion-dollar outcome. Some VCs self-select out on hearing that, which he considers entirely correct — that’s not the company for them, and that’s fine.

Key Takeaways

• Broker data isn’t inaccurate — it’s fragmented across systems that were each right at a single point in time and never reconciled afterward
• Data that can be purchased shouldn’t be manually entered; every avoidable manual field is a future correction waiting to happen
• Smaller brokerages often have cleaner data than large ones, because growth-by-acquisition at scale creates unreconciled data debt that takes hundreds of millions to fix
• The renewal friction problem is a design flaw, not a market reality — brokers avoid shopping the market because the process itself is punishing, not because the incumbent is genuinely best
• Automating commoditized-product pricing resequences work insurers already do with data; it doesn’t replace underwriting judgment on complex risk
• Claims data still runs on PDF-to-Excel manual re-entry at scale, creating both accuracy risk and lost placement-strategy insight
• Chasing a defined, smaller problem — and being honest that the company may never be a billion-dollar outcome — is a legitimate and durable way to build in this industry