Underwriters today have access to more information than at any point in the history of the profession: internal loss and exposure data, third-party data feeds, external scans, news, financial filings, and inspection reports. Yet many underwriters will tell you they feel less certain at the point of decision, not more.
The problem isn’t a shortage of data. It’s that relevant information arrives in the wrong place, at the wrong time, or buried under information that doesn’t matter. A prior decline sits in a different system. An accumulation concern is in a portfolio report nobody opens during the quote. An external scan produces forty findings, of which two actually bear on the risk.
A signal layer fixes this by surfacing the few things that should influence this decision, at the moment it’s made, with the evidence behind each one. This post covers how to design one that underwriters will trust and keep using.
Data, information, and signals
It helps to be precise:
- Data is everything available: application fields, loss runs, third-party records, scan results.
- Information is data organized around the risk: a structured submission, a loss summary, an exposure schedule.
- A signal is a specific finding that should change what the underwriter does or how they price, accompanied by the reason and the source.
“Insured had three water-damage claims in the last five years at the same location” is a signal. A table of every claim in the insured’s history is information. Both are useful, but only the first earns a place at the top of the screen.
What makes a good signal
Before adding a signal to the underwriting view, test it against five criteria:
- Relevant. It relates to the line, the coverage, and the decision being made. A signal that matters for property may be noise for professional liability.
- Material. It would plausibly change acceptance, terms, conditions, or price.
- Timely. It reflects the current state of the risk, not a snapshot from two years ago.
- Sourced. The underwriter can see where it came from and inspect the underlying evidence.
- Actionable. There’s a clear response: ask a question, add a condition, refer, adjust price, or decline.
Signals that fail these tests belong in the background information, not in the underwriter’s line of sight.
Where signals come from
Most useful signals fall into four groups.
Internal history. Prior submissions, declines, quotes, and claims for the insured or related entities. Prior broker relationships. Previous conditions and whether they were met. This is often the most valuable and most under-used source, because it lives in different systems.
Portfolio context. How this risk interacts with what you already hold: accumulations by location, port, vessel, cloud provider, or industry, and proximity to appetite and capacity limits. Underwriters rarely see this during the quote unless someone surfaces it.
External data. Third-party information such as property characteristics, corporate and financial records, sanctions and watchlists, and line-specific sources. It should always be used within the terms of the data licence and your rules on permitted rating factors.
Consistency checks. Conflicts within the submission itself: revenue that differs between the application and the financials, security controls attested on a questionnaire that external evidence doesn’t support, values on a schedule that don’t add up to the stated total.
Signals in specialty lines
The right signals vary sharply by line. Two examples:
Marine cargo. Useful signals include commodity type and its susceptibility to particular perils, packing and stowage arrangements, routes that pass through areas of elevated risk, storage periods at ports or warehouses, accumulation of insured values at the same port or facility, and loss history by commodity and route. A cargo underwriter who sees an accumulation warning for a port where the book is already heavily concentrated can act on it immediately, adjusting line size or requiring conditions.
Cyber. Useful signals include gaps between attested controls (such as multi-factor authentication, backups, and endpoint protection) and what can be observed externally, exposed services, the insured’s dependence on particular technology or service providers (which also matters for portfolio accumulation), previous incidents, and the sector’s current threat environment. A short list of material control gaps, each with evidence, is far more useful than a full scan report.
Avoiding alert fatigue
The quickest way to lose underwriters’ trust is to show them too many signals, too many of which turn out to be irrelevant. Practical safeguards:
- Rank and cap. Show the most material signals first, and limit how many appear prominently. Everything else is one click away.
- Tune thresholds with underwriters. Decide together what level of loss frequency, accumulation, or control gap merits a signal.
- Suppress what’s already handled. If a known issue is addressed by an existing condition, don’t raise it again at every renewal without a change.
- Explain every signal in plain language, with a link to the underlying evidence.
Close the feedback loop
Every signal should be measurable. Let underwriters mark signals as useful, not relevant, or wrong, and track which signals actually change outcomes: questions asked, conditions added, prices adjusted, referrals made, risks declined. Signals that never influence decisions should be retuned or retired. Signals that underwriters frequently flag as wrong point to data-quality problems worth fixing.
Over time, that loop also shows where losses later emerged on risks where a signal was shown but not acted upon, which is valuable for underwriting guidance and training.
Governance
Signals influence underwriting decisions, so they need the same discipline as any other decision input. Document each signal’s logic, source, and owner. Check that signals don’t act as proxies for prohibited factors. Record which signals were shown on each decision and how the underwriter responded. That record supports audits, capacity-provider reviews, and your own portfolio analysis.
Illustrative example: an MGA writing cyber for small and mid-sized businesses might replace the full external scan attachment with three ranked signals on the quote screen: any material gap between attested and observed controls, any prior incident, and any dependency on a provider where the book is already concentrated. Underwriters review fewer pages and ask sharper questions, and every quote records what was shown.
What to measure
- Signal precision: the share of signals underwriters rate as useful
- Decision influence: the share of signals that led to an action
- Time to decision for submissions with and without the signal layer
- Referral quality: whether referrals arrive with the relevant signals attached
- Loss outcomes over time for risks where signals were acted on versus not
Clarity at the moment that counts
Underwriting judgment is the value. A signal layer exists to protect it, clearing away the noise so that when an underwriter makes a decision, the few things that matter are right in front of them, with the evidence to back them up.
See what your underwriters are missing at the point of decision. Book the Aketud Underwriting Workflow Assessment to review the signals your underwriters see today, find the ones buried in other systems, and design a signal layer for your lines.
