Insurance Agency Management Software Problems: The 7 That Cost Retention and Commission, and How to Avoid Them

Why does the renewal workflow always get scoped as a report?

Because that is how it is described. Somebody asks for a renewal dashboard, the developer builds a list of expiring policies with filters, and it demos beautifully. Then the agency uses it for a month and goes back to Excel, because a list is not a workflow.

A renewal is a sequence of decisions, not a date. Review the account. Check premium change against expiring and check claims activity. Decide to hold or remarket. Assign the remarket to a producer with a deadline. Log the outcome, including the outcome where you decided to do nothing and why. Your incumbent gives you an expiration report and an activity log, which covers the first and last steps and none of the middle.

The fix is to model the pipeline rather than the list. Every policy crossing 120 days to expiration enters a stage automatically. Premium change comes off the IVANS download rather than a human comparison. Claims activity attaches to the account. Rules flag risk: an increase above fifteen percent, a monoline account that should be rounded out, a carrier appetite shift. Critically, no account can exit the pipeline without a logged decision, and that log doubles as your evidence that remarketing was considered when an errors and omissions claim arrives two years later.

Test any proposal against one question. If an account manager is out for a week, does the system know which of their accounts went unattended, or does that only surface when the insured calls? A report cannot answer it. A pipeline with owners and stage ages can.

What goes wrong when you migrate clients, policies and attachments out of AMS360 or Epic?

Three things, and none of them are the extraction itself.

The first is duplicate clients. Agencies that grew by acquisition carry the same insured under three records in three offices, sometimes under slightly different legal names, and both offices have mailed that client a renewal letter without noticing. Migration is where this becomes visible, and merging is a business decision with commission implications rather than a technical one. Somebody has to decide which producer owns the merged record.

The second is policy coding. Two offices that joined the agency at different times will have coded the same line of business differently, used different transaction types for endorsements, and disagreed about whether a rewrite is a new policy or a continuation. Any retention number you compute across those books is wrong until the coding is reconciled, and the reconciliation is manual.

The third is attachments. Documents export, but the links between a document and the specific policy term it belongs to are frequently weaker than the interface suggested. A certificate that was attached to an account rather than a policy loses most of its meaning once the account has five policies.

The pattern that works is unglamorous. Script the extraction so it can be rerun, not performed once. Rehearse the full load at least three times against production volumes. Then run one complete renewal cycle in parallel, comparing invoice by invoice and expiration list by expiration list, before anything is switched off. Agencies that compress this step are the ones that discover a coding problem in month three, after the old system is gone.

Why do IVANS downloads and carrier feeds break after launch?

Because carrier data is not a stable interface, and a build that assumes it is will start failing quietly in month two.

AL3 is a standard in the sense that everyone implements it, and it is a negotiation in the sense that carriers implement it differently. One carrier sends an endorsement as a full policy replacement. Another changes a code without notice. A third sends a malformed transaction after a system upgrade on their side. If your parser rejects a batch and nobody is watching, policies quietly stop updating and your renewal pipeline is running on stale premium data, which is worse than no data because it looks current.

The design decisions that prevent this are all defensive. Every inbound transaction is stored raw before it is parsed, so a parsing bug can be replayed rather than reconstructed. Transactions that fail to match an existing policy go to an exception queue with the raw payload attached, never to a log file nobody reads. Match on more than the policy number, because policy numbers change at rewrite. And alert on volume, since the most common real world failure is not an error message but a carrier that simply stopped sending.

Ask any prospective developer what they did the last time a carrier sent malformed AL3. If they have not had that experience, you will be paying for them to have it.

What happens when the errors and omissions trail is not covered?

You find out during a claim, which is the worst possible moment to discover a gap in your own records.

The defence in most agency claims rests on showing what was recommended, what the insured declined, and when. That means an immutable record: who advised what, on what date, with what coverage options presented, and what the insured chose. Systems that let a note be edited after the fact, or a decision to be backdated, are not producing evidence. They are producing a document that opposing counsel will enjoy.

The specific technical commitment is append only history for coverage recommendations, declinations and communications. Corrections are new entries that reference the original rather than overwrites. Timestamps come from the server, not the workstation. Exports are complete and readable by someone who does not use your software.

Two related obligations belong in the same architecture conversation rather than in a later hardening pass. State insurance data security laws modelled on the National Association of Insurance Commissioners model law impose specific requirements on how customer data is protected and how incidents are handled. If you write business in New York, the Department of Financial Services cybersecurity regulation applies to you directly. Producer licence tracking and trust account separation for agency bill business are the other two that get deferred and should not be. A partner who proposes adding compliance later is telling you they have not built in this industry.

Should you build custom or configure what you already own?

Under about fifteen seats in a single location with a standard personal lines book, do not build. The per seat maths still works at that size, HawkSoft or the EZLynx Management System costs far less than Applied Epic, and your workflows are close enough to generic that customisation buys very little. We say this to agencies regularly and mean it.

There is a second group who should also stop. If your renewals fail because nobody owns them rather than because the software cannot express them, new software will not fix it. Assign the book, set the review cadence, and see what is left after a quarter. The agencies that get the most from a build are the ones who have already designed the process by hand.

The build case is specific. Forty or more seats and a six figure annual management system bill. Renewals, commissions or certificates already running in spreadsheets and email around the incumbent, which means your team has designed the software and is executing it manually. Multiple offices or an acquisition habit, so consolidation is a recurring cost rather than a one off. Or a programme and niche book, in trucking, contractors or benefits, where your process is the edge and a generic system flattens it.

Our standing recommendation for that profile is to build the workflow layer on top of your existing data first, keep Epic or AMS360 as the system of record, run it for two quarters, and only then decide whether replacing the core is worth it. Most agencies find the workflow layer captures the bulk of the value at a fraction of the cost.

How do hidden costs get into the quote?

Commission statement parsing is the most reliable surprise. Forty carriers means forty statement formats, and each parser is real work: portal downloads for surplus lines, spreadsheets from nationals, and PDFs from regional carriers whose layout changes when they redesign their letterhead. A quote that says commission reconciliation without naming which carriers are automated in phase one is not a quote, it is an aspiration. Name them, and agree what happens to the rest.

ACORD form generation is the second. There are many forms and each has its own field mapping and version history. The ACORD 25 certificate alone is not the job if your contractors book needs additional insured endorsements and manuscript wording review.

Premium trust accounting is the third, and it is architectural. If you handle agency bill money, the separation requirements shape the ledger design and cannot be retrofitted cheaply.

The fourth is producer compensation. Tiered overrides, split changes that take effect mid year, and house accounts each add rules, and effective dating those rules properly is what stops the disputes that currently arrive every month. Scope it explicitly or it will arrive as a change request during user acceptance testing.

What separates a build that works from one that fails here?

Make them whiteboard the policy data model before you sign. Policies against lines against transactions. Effective dated endorsements. Agency bill against direct bill. Cancellations, reinstatements and rewrites. If they sketch a policy as one row with a premium column, the build collapses at the first endorsement and everything downstream inherits the error.

Demand named carrier connectivity experience. IVANS enrolment, AL3 or ACORD XML parsing, and download reconciliation against existing policies are the three that matter, and there is no substitute for having done them.

Inspect the migration plan as a document, not a paragraph. How many rehearsal runs, at what data volume, and how long the dual run lasts. One full renewal cycle in parallel is the safe answer and anything shorter should be justified.

Settle ownership before kickoff. You should hold the repository, the cloud accounts and the unrestricted right to hire another firm. At Digital Heroes the agency owns the code from the first commit. Your renewal rules and commission structures are competitive assets, and renting them back from a vendor is the situation you left the incumbent to escape.