A certificate of insurance is one of the simplest documents in commercial lines. The customer’s policy already exists. The carrier already issued it. The data — limits, endorsements, additional insureds, certificate holder — is sitting in your AMS, where you put it when you bound the policy.
And yet every COI request is still, at most agencies, a human task.
The anatomy of one COI request
A commercial customer emails the service inbox at 9:14am: “Hi, can you send me a certificate naming my new landlord as additional insured? Address attached. Thanks.”
Here’s what has to happen next:
- Someone reads the email and decides it’s a COI request.
- They find the customer in the AMS by name or email address.
- If the customer has multiple policies, they identify which one the request is for. (Commercial GL? Commercial auto? Both?)
- They confirm the holder details and the wording the customer wants — sometimes by going back to the email, sometimes by guessing.
- They generate the certificate. In most AMSes, this is a templated form-fill: paste the holder, paste the address, check the endorsements, save.
- They email it back to the requester, often with the carrier as a CC.
Best case: five minutes. Average case: ten. Bad case (the customer’s account has multiple policies, the request is ambiguous, the policy is up for renewal): thirty.
Multiply that by however many COI requests your agency takes in a week. For a typical mid-size commercial book, that’s somewhere between twenty and a hundred a week.
The hidden costs nobody talks about
The minutes are the visible cost. There are three others.
Latency cost. Most COI requests have a deadline attached — the customer needs the cert today because the landlord wants it before the lease signs at 4pm. If your service team is buried, the certificate arrives at 5:30. The customer is unhappy. They remember next time. If they have a complex book and a friend at another agency, they shop.
Producer drag cost. A non-trivial fraction of COI requests get bounced from the service inbox to a producer because they involve “interpretation” — usually a wording question or an endorsement question that the AI-shy CSR doesn’t want to get wrong. Now the producer is mid-quote-call and an email pulls them out. The quote suffers.
Error cost. Manual COI handling makes manual COI errors. Wrong holder name. Wrong limit. Wrong endorsement listed. Most are caught before they’re sent. The ones that aren’t either embarrass you with the customer or — if it’s bad enough — surface in a claim later as a coverage dispute.
The minutes are real. The latency and the drag and the error tail are bigger.
What automation changes
For agencies running Otto Agent, the same incoming email triggers a different sequence:
- Otto reads the email as soon as it lands in the shared mailbox.
- Otto identifies the customer in the AMS by email or phone number.
- If the customer has multiple profiles (personal + business, or multiple businesses), Otto identifies which one the request is for and pulls the correct policy.
- Otto generates the certificate from the policy data.
- Otto sends it back to the requester.
End-to-end is typically three to five minutes. The producer never sees it. The CSR never sees it. The customer doesn’t notice anything except that the certificate showed up faster than usual.
Two things deserve a separate paragraph, because they’re the ones service-team leads ask about first:
Same-day-expiring policies. Otto won’t auto-issue a certificate for a policy expiring the same day. Those are too risky for full automation — they get routed to your service team for manual review. The control stays with humans for the edge cases that matter.
Review-copy routing during rollout. When you first turn Otto on, you can route every generated certificate as a review copy to the service team before it goes to the requester. Your team validates the output, builds confidence, and then flips it to fully automated when they’re ready. No big-bang trust ask.
How to think about the ROI
If your agency is running 40 COI requests a week at 10 minutes apiece, that’s about 350 hours of CSR or producer time per year on something that, mechanically, is just data movement. The agencies we’ve talked to value that time anywhere between $30 and $80 an hour depending on who’s actually doing it.
That’s between $10,000 and $28,000 a year, sitting inside the COI inbox.
It’s not the whole reason to automate. The latency and error tail are arguably bigger. But it’s the number that makes the case to a partner who wants to see a spreadsheet.
The honest part
Automating COIs isn’t sexy. It’s not the thing you’ll lead with at the next agency leadership meeting. It is, however, the kind of unglamorous workflow that produces real recurring time savings without changing how your team works.
If you want to see what it looks like on a real agency’s mailbox, book a 15-minute call and we’ll walk through Otto on a live inbox.
Aiste Buinauskaite
Head of Business Development