Renewal automation that agencies actually keep using

A missed renewal is not one lost policy. It is the commission on that policy, the account it was attached to, and usually the other lines that client had with you.

That asymmetry is what makes renewals worth automating before anything else in an agency. A quote that does not close costs you the time you spent on it. A renewal that lapses costs you an existing relationship that was already paying, and it does so quietly — nobody sends a notification saying a client left. You find out when the commission statement is short, which is months later.

The other thing that makes renewals a good candidate is that the work is almost entirely knowable in advance. You know the expiration date the day the policy is issued. Nothing about the calendar is a surprise. That is exactly the shape of problem a rule handles well and a person handles badly, because a person has to remember and a rule does not.

Three marks before expiration, and each one exists because something stops being possible after it.

90 days is early enough to notice without being urgent. Nothing has to happen yet; the point is that the renewal stops being invisible. An agency that first looks at a renewal at 30 days is not doing renewal work, it is doing damage control.

60 days is set by the carrier, not by you. Loss runs and exposure information take time to come back, and if you request them at 45 days they arrive too late to inform a remarket. This mark exists because of someone else's turnaround time.

45 days is the last point at which the outcome can still change. After it you can process a renewal but you cannot really shop it, and the conversation with the insured stops being "what do you need this year" and becomes "sign this". Past 45 days the window is running the desk.

The most common way renewal automation goes wrong is trying to automate judgement. The line worth holding is simple: rules handle what is known in advance, people handle what is not.

Belongs to rules:

• Flagging the renewal when the window opens.
• Requesting loss runs on a schedule tied to the expiration date.
• Routing the file to the producer who owns the account.
• Listing which certificate holders will need a reissue.
• Updating the commission schedule once the renewal binds.

Belongs to a person:

• Whether this account should be remarketed at all.
• Which carriers to approach, given appetite and history.
• What to tell an insured whose exposures changed.
• Whether a premium increase is worth defending or negotiating.

Automation that stays on the first list gets trusted. Automation that drifts into the second starts producing outputs the agency has to check, and checking a machine is more work than doing it yourself — which is the beginning of the end for that feature.

Concretely, this is the work one commercial policy generates across a renewal cycle, and when each piece opens.

• 90 days — 90-day renewal reminder sent to the insured
• 60 days — 60-day reminder sent to the insured
• 60 days — Loss runs requested from the carrier
• 45 days — 45-day reminder sent to the insured
• 45 days — Remarket decision · renewal meeting
• 45 days — Certificate holders to update
• At renewal — Renewal bound · new term issued
• At renewal — Certificates reissued to holders
• At renewal — Commission schedule updated

Nine items for one policy. An agency with 300 commercial policies is generating roughly 2,700 discrete renewal actions a year, spread unevenly across the calendar. That volume is why "we have a good process" and "we have a reliable process" are different claims: the process is usually fine, and it is the remembering that fails under load.

Notice also that the last three items happen after the renewal binds. Reissuing certificates and updating the commission schedule are the steps most often dropped, because by then the renewal feels done. They are also the steps that cause problems later — a holder with an expired certificate, or a commission that never reconciles.

Renewal automation usually is not removed. It gets ignored, which is worse, because the agency keeps believing it is covered.

Three failure modes cause almost all of it:

• Alerts with no action attached. A notification that says "policy expiring" and nothing else moves the work to whoever reads it. After a few weeks people stop reading.
• Firing on the wrong data. If the expiration dates in the system are unreliable — a common consequence of a bad migration — the automation produces false alarms, and false alarms train people to dismiss it.
• No visible ownership. Work that routes to "the agency" routes to nobody. Every automated item needs a name attached, which is why the producer/servicer split matters more than it sounds.

The common thread is that automation earns trust by being specific and loses it by being noisy. A system that opens three precise tasks is more useful than one that sends thirty reminders.

Two numbers tell you whether renewal automation is working, and neither is "alerts sent".

• Retention rate by line. The outcome that matters. Track it per line of business, because a blended number hides the department that is losing accounts.
• Share of renewals worked before 45 days. The leading indicator. If this is low, retention will fall later regardless of what the automation reports today.

A third one is worth watching if you can get it: how many certificates were reissued after a renewal bound, against how many holders existed on the expiring policy. A gap there is future work arriving as an emergency.

In Agency Nexus, the rule part is a reminder to the insured at each of the three marks, by SMS or email; the rest of the list above stays with the agency. The renewal automation page says exactly what runs on its own and what does not.