Submission volume is rising, underwriting capacity is not

Submission Volume Is Rising, Underwriting Capacity Is Not

A regional MGA placed 847 submissions across their carrier panel in Q3 2025. By Q1 2026, that number reached 1,104. Nobody hired more underwriters. Nobody expanded the team.

This is no longer an outlier story. It is the market condition.

Why This Moment Matters

The P&C market is entering a correction phase. Rate relief in property, ranging from 10-30% depending on territory and class, has shifted broker behavior. When premiums soften, brokers shop. They shop more. They send submissions to carriers and MGAs they may not have approached in the hard market, knowing that competition will drive better terms.

The result is structural: submission volume has decoupled from headcount.

Meanwhile, the earnings profile remains uneven. Workers' compensation has delivered 12 consecutive years of profitability, with combined ratios holding below 100%. Commercial auto continues to bleed, with net underwriting losses exceeding $10 billion over the past two years. Social inflation, driven by higher jury verdicts and increased claims costs, compounds the challenge across liability classes.

This creates a staffing paradox. Underwriting resources are finite. They migrate toward profitable lines. Commercial auto, despite being broken, still demands bandwidth because the volume is there. Property demands attention because rate changes happen weekly. Workers' comp, profitable as it is, often receives fewer specialists because the book performs. Capacity becomes allocated, not multiplied.

Direct premiums written are expected to grow 4-5% in 2026. Underwriting teams at most carriers remain flat or contracted slightly.

The Capacity Bottleneck Is Not What Most Assume

When underwriting leaders talk about capacity constraints, the conversation usually goes to headcount. "We need more underwriters." The implication: hire your way to a solution.

That framing misses the actual problem.

A typical commercial lines underwriter processes between 3-7 submissions per day, depending on complexity and the quality of intake documentation. That sounds reasonable. But when each submission requires 20-30 minutes of manual document review, data extraction, and file organization before the underwriting decision even begins, the math breaks. Ten hours spent on extraction and intake is ten hours not spent on risk selection, renewal strategy, or loss analysis.

Capacity is not a headcount problem. It is a time-allocation problem.

Many carriers and MGAs have invested in submission intake automation tools, but adoption remains uneven. Some systems are tied to specific document templates. Others require pre-processing of submissions before they extract data. A few are actually learning systems that improve with use, but these remain uncommon.

The result is that underwriters still spend disproportionate time on clerical work, not expert judgment.

What Changed, and Why It Matters Now

Three factors have created this inflection point simultaneously.

First, broker behavior has shifted measurably. In a hard market, brokers develop deep relationships with a few preferred carriers and concentrate submissions. As margins compress and rate relief materializes, that behavior reverses. The same broker now sends the same risk to five carriers instead of two. From a carrier perspective, this means more inbound volume with no corresponding increase in the size of your underwriting team.

Second, document volume per submission has increased. Ten years ago, a commercial package might include five documents: declarations, loss runs, financial statements, prior loss history, and a one-page application. Today, brokers expect carriers to review certificates of insurance, detailed operations manuals, OSHA records, audit reports, social media profiles for reputational risk, and board minutes. The submissions themselves have become denser. The expectation for underwriter review has expanded. Time spent on intake has nearly doubled.

Third, underwriting talent remains scarce. The market has not produced a cohort of newly trained underwriters proportional to demand. Technical skills like loss run analysis require years to develop. Hiring external talent to backfill capacity is expensive and inefficient; onboarding takes months. Most carriers have responded by keeping experienced underwriters longer, accepting higher turnover in junior roles, and centralizing high-touch decisions. This consolidation is rational from a risk perspective. It is disastrous for throughput.

The gap between submission volume and underwriting capacity is not cyclical. It is structural.

How Pibit.AI Thinks About This Problem

At Pibit.AI, we work directly with 40+ commercial P&C carriers and MGAs. We have watched this tension play out across carriers of different sizes and geographies. The consistent thread: organizations that succeeded in 2025-2026 did not hire their way out of the problem. They engineered their way out.

Specifically, they automated the parts of underwriting that do not require expert judgment, freeing underwriters to do what they actually add value on: risk selection, relationship management, and strategic decision-making.

This is not replacement. It is augmentation. The best underwriting teams are still staffed with senior talent. The difference is that senior talent spends its time on decisions, not data entry.

Our CURE™ platform extracts data from any document format, regardless of format or broker variation. It builds a structured submission in minutes, not hours. Underwriters see clean, organized information before they ever open a file. The extraction is template-agnostic and does not require pre-processing.

The outcome is measurable: carriers using our platform process submissions 85% faster. Underwriters generate 32% more gross written premium per person. And the decisions themselves are more consistent; our clients report 700 basis points of loss ratio improvement through better risk selection accuracy.

But the underlying insight matters more than the platform itself: you do not fix capacity constraints by hiring more. You fix them by reclaiming time spent on non-expert work.

The Math Is Brutal, But Clear

Consider a mid-sized carrier with 35 underwriters processing roughly 200 submissions per week. At current efficiency, that is 5.7 submissions per underwriter per week, or just over one per day. Each submission takes roughly two hours from intake to submission in the underwriting system. That is the current baseline.

Now assume submission volume increases 25% (conservative, given market dynamics). The carrier could hire 9 additional underwriters to maintain processing speed. Or, it could automate intake and extraction, cutting the time per submission from 120 minutes to 20 minutes. The second path frees up roughly 1,166 hours per month. That is equivalent to 5-6 additional underwriters, with none of the onboarding, retention, or compensation cost.

The gap between submission volume and capacity-building efforts is not being closed by headcount. It is being closed by process automation, and organizations that have adopted this approach are pulling ahead.

Consider also the competitive effect. A carrier that turns a submission in 48 hours (vs. 5 days) wins the deal. A carrier that returns submission decisions with clear feedback wins renewals. A carrier that handles 25% more volume with the same team wins market share in a soft market where every piece of growth matters.

Why AI in Underwriting Matters Here

Machine learning and natural language processing have matured enough to extract structured data from unstructured documents at scale. This was not true five years ago. Accuracy was the bottleneck. That bottleneck is gone.

Modern extraction systems learn from corrections. If an underwriter corrects an extracted field, the model improves. Over time, accuracy approaches 99-100% for standard fields. For complex items like policy conditions or loss narrative, systems now validate against multiple documents and flag ambiguities, rather than guessing.

The practical implication: underwriters trust the extraction. They do not need to re-verify every field. They review the decision-critical items and move forward.

This is the "Centaur Underwriter" model: the human expert augmented by AI, not replaced. The underwriter's judgment on risk appetite, relationship factors, and strategic fit remains irreplaceable. But the underwriter's time on data gathering and manual entry is recovered.

For organizations grappling with underwriting challenges, this rebalance is no longer optional. It is competitive necessity.

Practical Steps: Three Things to Test Now

First, audit where your underwriters actually spend time. Track a sample of submissions from intake to decision. Capture time spent on document review, data entry, system navigation, and actual risk selection. Most organizations discover that 40-60% of time is spent on non-expert work. That is the addressable opportunity.

Second, evaluate whether your current submission intake system is template-agnostic. If it works only with your standard forms, it will fail as submission volume diversifies. As brokers shop more widely, submissions will arrive in varied formats. Your system should handle that variation without manual intervention.

Third, measure the true cost of submission processing time. Calculate fully loaded cost per submission. Then calculate what a 50% reduction in intake time would mean for your underwriting cost structure. That math often justifies investment in automation immediately.

The Window Is Open, But Not Forever

Soft markets do not last indefinitely. Rate correction will eventually stabilize. When it does, submission volume will moderate. The carriers that have adapted their processes in 2026 will be operationally more efficient when the hard market returns. Those that rely on hiring alone will face cost structure problems when volume declines.

This is the inflection point. Submission volume is rising. Underwriting capacity is not. The gap closes through process, not people.

If you are an underwriting leader evaluating how to handle 2026, this is the operational question that matters most: how do you process 25% more submissions with the team you have today?

The answer is embedded in how you approach intake, extraction, and data orchestration. Not in how many bodies you add to the bench.