Every property and casualty brokerage runs on submissions. A client calls, a CSR gathers data, forms get filled, quotes come back, proposals go out. It sounds simple. In practice, it's where brokerages bleed the most time and accumulate the most risk.

We've spent significant time embedded inside mid-market brokerages — watching CSRs work, timing workflows, cataloguing where things break. What we consistently find is a system held together by institutional memory, browser tabs, and hope.

The Data-Gathering Problem

For a single home submission, a CSR typically pulls data from multiple separate sources — property risk data from one portal, tax and parcel records from another — then manually transfers that data into a Dwelling Replacement Cost Estimator. This process commonly takes 15 to 30 minutes per form. The portals themselves create additional friction with session limits, rate caps, and inconsistent UX.

When we built Compass, our submissions platform, we integrated directly with these data sources. The result is a near-complete auto-fill rate on home DRCEs. Data that used to require 20 minutes of copy-paste now populates in seconds. Multiply that across an agency's annual submission volume and the time savings become material.

The Coverage Gap Problem

The more dangerous issue isn't speed — it's what gets missed. Coverage gaps are the primary driver of E&O claims in the brokerage channel, and the consultancies who do post-mortem reviews on agency E&O exposure consistently surface the same root causes: weak procedures manuals, inconsistent documentation, unclear producer-vs-CSR responsibility, and missing disclaimers at exactly the points where claims later get contested.

The pattern beneath those root causes is straightforward: something changes — a carrier modifies coverage at renewal, a client modifies their exposure, an underwriter quietly drops a coverage line when moving carriers — and the change doesn't get detected or documented. By the time a claim is filed, the paper trail doesn't exist.

Compass runs an AI-powered coverage analysis on every submission. The vast majority of analyses surface at least one issue worth a human review. The point isn't that the AI is catching everything — it's that the conversation now happens, every time, because a system is asking the question even when humans forget to.

The Proposal Problem

Even after data is gathered and quotes are obtained, the last mile is surprisingly broken. Most brokerages have no standardized proposal format. E-signatures are handled through a patchwork of methods — in-person signing, generic PDF tools with manually placed fields, print-scan-fax loops. There's no consistent coverage summary, no expiration date, no written confirmation when coverage is declined.

This matters because the proposal is the primary E&O defense document. If a client later claims they weren't offered a coverage option, the proposal is what proves otherwise — but only if it exists, is standardized, and was actually signed.

Compass generates proposal packets with standardized coverage summaries, AI-written introduction letters, and integrated signature workflows. Beyond the efficiency gain, this is a paper trail that didn't exist before.

What This Means for the Industry

The brokerage channel processes billions in premium annually through workflows that haven't fundamentally changed in decades. The tools have gotten prettier — cloud-based management systems, carrier portals with better UIs — but the core process remains: a human copying data between systems, a human remembering to check coverage, a human assembling a proposal by hand.

The opportunity isn't to replace the human. Insurance is a relationship business, and the judgment calls that producers and CSRs make every day are genuinely valuable. The opportunity is to eliminate the mechanical work that surrounds those judgment calls — so that human time goes toward advising clients, not filling forms.

The workflow problems we find inside brokerages aren't unique to any one market or agency size. They're structural to how the industry operates. The question isn't whether these workflows will be automated. It's whether the automation will be built by people who've actually sat in a brokerage and watched the work happen.