Why patching evidence now moves your quote more than your answers do
If your renewal came back with a surcharge, a ransomware sublimit or a condition like "bind subject to remediation within 30 days," patching is the likely cause. Underwriters rarely take the application at face value anymore. Many carriers run outside-in scans of your domains and IP ranges before quoting, then compare the results with what you told them.
Patch management requirements for cyber insurance usually reach you in three forms: application questions about remediation timelines and unsupported software, scan findings tied to specific CVEs, and subjectivities you must clear before or after binding. Strictness varies by market. Carriers in London and the US tend to scan hardest, while insurers in parts of Asia, the Middle East and Latin America may lean more on questionnaires. Regulation adds its own layer. PCI DSS v4.x expects critical security patches within a short, defined window (one month for many cases), and the EU's NIS2 and DORA regimes require documented vulnerability handling for the entities they cover.
The seven gaps below are the ones underwriters tend to flag. Each comes with a fix and the evidence that usually gets a penalty removed.
Looking to streamline vendor RFP selection?
Access pre-negotiated tier discounts through our independent institutional network.
Keeping end-of-life servers in production without a plan
The mistake: leaving an old Windows Server, an unsupported Linux distribution or a retired database in place because a line-of-business app depends on it, then answering "no" to the unsupported-software question. Underwriters ask bluntly because once a system stops getting patches, every new flaw in it stays open for good.
What to do instead: pick one of three paths for each system and write it down.
- Retire or migrate it, with a dated decommission plan.
- Buy extended support, such as Microsoft Extended Security Updates, Red Hat's extended life cycle offerings, or a third-party provider like TuxCare. ESU is typically priced per core or per server and tends to rise each year. For a small server estate, expect roughly a few thousand to tens of thousands of USD a year. These are indicative 2026 ranges that vary by contract and region, so confirm current pricing with the vendor.
- Isolate it: no internet exposure, a segmented VLAN, tight firewall rules and no domain admin logons.
Hypothetical example: a 150-person manufacturer runs its production software on two end-of-life servers. If it shows the underwriter an extended-support invoice, a network diagram proving the servers can't reach the internet, and a migration date nine months out, it usually gets a far better hearing than if it just says "we have legacy systems."
Evidence: an asset register flagging every EOL system, proof of extended support or isolation, and the decommission date.
Letting VPNs, firewalls and edge appliances fall behind on firmware
The mistake: putting VPN concentrators, firewalls and remote-access gateways on the same monthly cycle as desktops. Underwriters care about these internet-facing devices most. Products from Fortinet, Ivanti, Citrix, Palo Alto Networks and SonicWall have all had actively exploited flaws in recent years, and an unpatched edge device is one of the fastest routes to a decline.
What to do instead: treat edge devices as their own tier with their own rules.
- Subscribe to each vendor's security advisories and to the CISA Known Exploited Vulnerabilities (KEV) catalog. A US agency publishes it, but carriers worldwide use it.
- Patch actively exploited edge flaws within 48 to 72 hours. If you can't, apply the vendor's mitigation or take the service offline.
- After patching a device that was exploitable while exposed, run the vendor's integrity or compromise-check tool if one exists. A patch doesn't evict an attacker who's already inside.
- Close exposed RDP and require MFA on every remote-access path.
Evidence: the current firmware version, timestamped change tickets, the integrity-check output and a clean external rescan. Ask your broker to get the carrier's scan re-run once you're done.
Running without a written critical vulnerability remediation SLA
The mistake: saying "we patch promptly" with no policy behind it. Without a defined critical vulnerability remediation SLA, an underwriter can't tell whether a critical flaw stays open for a week or a quarter.
What to do instead: adopt a short management-approved policy with severity tiers and a rule that known-exploited flaws jump the queue. The timelines below are common market expectations, not a universal standard. Check them against your carrier's wording and any sector rules that apply to you.
| Category | Typical target | Notes |
|---|---|---|
| Known exploited, internet-facing | 48-72 hours | Mitigate or isolate if no patch exists |
| Critical (CVSS 9.0+) | 7-14 days | Shorter if exposed to the internet |
| High (CVSS 7.0-8.9) | 30 days | Roughly in line with PCI DSS's one-month expectation |
| Medium | 60-90 days | Bundle into regular cycles |
Evidence: the signed policy plus three to six months of metrics: the share of findings closed within SLA and mean time to remediate by severity. The metrics carry more weight than the policy because they show it's actually followed.
Looking to streamline vendor RFP selection?
Access pre-negotiated tier discounts through our independent institutional network.
Scanning only the assets you already know about
The mistake: pointing a scanner at the on-premises subnet and calling it vulnerability management. Cloud workloads, remote laptops, acquired subsidiaries and forgotten test servers go uncovered, and they're often exactly what the carrier's external scan finds.
What to do instead: for underwriting purposes, vulnerability management means proving coverage, not owning a scanner. Reconcile your asset inventory against what the scanner sees. Use agents or authenticated scans rather than unauthenticated network sweeps. Add external attack surface monitoring so you see what the carrier sees. Tenable, Qualys and Rapid7 are all available internationally. For mid-market firms, budget very roughly USD 20-60 per asset per year, depending on modules and volume. That's an indicative 2026 range, so confirm with the vendors or a local reseller.
Evidence: a coverage figure (assets scanned divided by assets in inventory), with a high target such as 95 percent or better, and an external attack surface report dated within 30 days of submission.
Patching the operating system and ignoring everything else
The mistake: showing clean Windows Update compliance while browsers, PDF readers, Java runtimes, file-transfer tools and hypervisors drift. Mass exploitation of managed file-transfer products and ransomware aimed at VMware ESXi have taught underwriters to ask about these layers directly.
What to do instead: automate third-party application patching with whatever your endpoint stack supports. Options include Microsoft Intune with a third-party catalog, ManageEngine, NinjaOne, Automox and Action1. Put hypervisors, storage controllers and network switches on a documented lifecycle with named owners. For OT or medical devices you can't patch, segmentation and monitoring serve as compensating controls, which leads to the next gap.
Evidence: patch compliance reports broken out by category (OS, third-party apps, hypervisors, network firmware) rather than one blended percentage.
Granting exceptions with no owner, expiry or compensating control
The mistake: leaving a system unpatched because "the vendor won't certify the update," with no record of it. Exceptions are normal. Undocumented ones look like neglect, and some policy wordings reduce coverage through coinsurance or sublimits when a loss traces back to a known vulnerability left open past a set period.
What to do instead: keep an exception register. For each entry, record the asset, the CVE, the business reason, the compensating controls (segmentation, EDR, disabled services, extra monitoring), the risk owner and an expiry date no more than 90 days out. Review it monthly and re-approve or close every entry.
Evidence: the register itself. A short, current register with named owners tells an underwriter you know your risk. No register suggests you don't.
Answering the application from memory instead of from data
The mistake: letting whoever fills in the renewal form write "critical patches applied within 14 days" when the carrier's scan shows a year-old critical CVE on your mail gateway. At best, the mismatch costs you credibility and pricing. At worst, it can give the insurer grounds to dispute a claim. Outcomes depend on jurisdiction: the UK's duty of fair presentation, civil-law disclosure rules in much of Europe and Latin America, and US state law each treat misstatements differently. A coverage lawyer or your broker can tell you which rules apply to you.
What to do instead: build the answers from an evidence pack: the remediation SLA, last quarter's SLA metrics, scanner coverage, an EOL register, the exception register and a fresh external scan. Have the CISO or IT lead sign off on the patch management answers before the CFO submits.
Your next step: 90 days before renewal, run your own external scan and ask your broker for the carrier's findings from last year. Fix any internet-facing critical or KEV-listed finding before anything else on this list. If you can't close it before submission, document the mitigation and a fix date. Disclosing a gap with a plan usually prices better than letting the carrier find it.