Manual insurance document review is risky because a certificate of insurance summarizes a policy without proving what it covers, and human reviewers working at volume tend to verify the summary rather than the policy. The predictable results are unrequested endorsements, misread entity names, missed expirations, unseen exclusions, undocumented decisions and delays that stall work.
Why is the certificate itself the first risk?
Every manual review starts from a document that cannot carry the weight placed on it. A certificate of insurance is informational only, confers no rights and cannot amend the underlying policy. It records what the broker believed the policies looked like on the day it was issued.
A manual process built around collecting certificates therefore verifies intent, not coverage. Limits, dates and checkboxes can all be correct while the policy behind them excludes the work, names a different entity, or grants a narrower endorsement than the contract required. Each of the risks below is a variation on that gap.
Which endorsements get missed in manual review?
The most common failure is accepting certificate wording in place of the endorsement. A line in the Description of Operations stating that the owner is an additional insured describes the broker’s understanding and does not itself grant coverage; the endorsement is the policy form that does. Requesting it means a second round trip to the broker, which busy teams skip.
When endorsements are collected, manual reviewers often stop at the first page. The schedule may be blank or name the wrong entity, the policy number may not match the certificate, the dates may belong to a prior term, or the form may be a specimen that was never issued. The same pattern applies to waivers of subrogation, which are required per policy line: a waiver evidenced on general liability does not cover workers’ compensation or auto, and each line has its own checkbox on the ACORD 25.
How do entity name errors survive review?
Coverage belongs to the exact legal entity on the policy. Contracts are signed with one entity, certificates are issued for another, and the two look alike: a trade name or DBA instead of the legal name, a holding company instead of the operating company, or a near-identical LLC set up for a different state or project.
Manual review reads names by recognition. The reviewer sees the company they contracted with and moves on. A one-letter difference survives hundreds of reviews because nobody decides to accept the wrong name; it never registers as a question. The same recognition error applies to the certificate holder box, which reviewers sometimes treat as protection even though a certificate holder is entitled to no benefits under any circumstances.
Why do expirations and cancellations slip through?
A certificate is accurate for one day. Policies renew, cancel and change mid-term, and a manual file has no way to know. Teams that track expirations in spreadsheets or inbox reminders depend on someone noticing the date, requesting the renewal certificate, and re-reviewing the endorsements that came with it.
Cancellation is worse. Most reviewers assume the “30 days notice” line on a certificate protects them. It does not: the ACORD 25 says notice will be delivered in accordance with the policy provisions, and in practice most insurers do not send notice unless a written contract or endorsement obligates them. Without a notice of cancellation endorsement on file, a vendor’s policy can lapse mid-project and nobody is told.
Why do exclusions never appear in a manual review?
Exclusions are the risk manual review is least equipped to catch, because they do not appear on the certificate at all. A general liability policy can exclude work above a certain height, residential projects over a unit count, particular materials, or specific classes of injury claims, and the certificate for that policy will show full limits, current dates and every endorsement attached.
Finding an exclusion means requesting the policy or its schedule of forms and reading the endorsements themselves, including ones with generic titles like Designated Work or Classification Limitation whose restrictions sit inside a schedule box. Manual COI review almost never reaches that document. The result is a compliant file and a policy that will not respond when the claim matches the excluded work.
What is the cost of having no audit trail?
Manual review leaves few records of why a decision was made. A waiver granted verbally, a requirement relaxed in an email, or an exception approved by someone who has since left the company all disappear into inboxes. At claim time, or during an internal audit, the question “who approved this and on what basis” has no documented answer.
The consequence is liability that cannot be defended and a compliance program that cannot be improved, because there is no data on where it fails.
How do review delays create their own risk?
When each certificate takes days to review and each deficiency takes days to resolve with the broker, teams face a choice between holding work and letting it start unverified. Under schedule pressure, the second option wins. Subcontractors mobilize before their coverage is confirmed, pay applications are held after the fact, and the compliance function becomes a bottleneck rather than a control.
The delay compounds the other risks: a slow process is also a process where the endorsement request is dropped, the renewal is missed and the exclusion is never read.
What does this mean for how review should work?
Each risk above has the same fix in principle: review the policy documents that decide coverage, compare them against the contract, and record the result. Doing that manually at the scale of a large project or a property portfolio is where the process breaks. Jones automates the collection and review of certificates and endorsements, checks each form against the contract requirements, and documents every gap and its resolution. How AI-assisted review compares with the manual process, and where a human is still needed, is covered in our comparison of AI-assisted and manual COI review.
FAQs
If the COI says compliant, why isn’t that enough?
Because the certificate only summarizes the policy. It confers no rights and cannot amend the coverage. The endorsements and the policy itself determine whether coverage applies, and a certificate can be accurate while the endorsement behind it is narrower than the contract requires or the policy excludes the work.
What is the most common manual review mistake?
Accepting wording on the certificate in place of the endorsement. Additional insured status, waivers of subrogation and notice of cancellation are all granted by endorsement, and the certificate can state them without the endorsement existing or without it matching the contract.
How do exclusions get past a compliant COI?
Exclusions live in the policy and its schedule of forms, not on the certificate. A policy can exclude specific work, project types or claim categories while the certificate shows full limits and every required endorsement. Only a review of the policy documents surfaces them.
Why does a missing audit trail matter?
At claim time or in an audit, you need to show what was reviewed, what was approved and why. Manual processes leave decisions in emails and memory, so exceptions cannot be defended and the program cannot be improved.
Jones reviews certificates and the endorsements behind them against contract requirements, flags every gap with instructions to fix it, and keeps an audit-ready record of each decision. See how the Jones Platform handles review.
