When an agent sends a commercial lines submission in 2026, the workflow that follows looks remarkably similar to what it looked like in 1995. The form arrives as a PDF, usually an ACORD 125 with supporting forms, and someone on the receiving end opens it, reads the fields, and types the information into a different system. The primary change since 1995 is that the document travels by email instead of by fax. The document itself, and the labor required to process it, are largely unchanged.
This was not the inevitable outcome of how technology evolved. ACORD, formed in 1970, accomplished something genuinely difficult: it persuaded competing insurance carriers to agree on a common set of forms. Before standardization, every carrier had its own application, its own fields, its own layout. Agents kept separate filing systems for each market they accessed. ACORD replaced that with a common language: the 125 for commercial lines, the 126 for general liability, the 140 for property, and eventually several hundred others covering specialized lines. By most measures the adoption was successful, and the vast majority of commercial submissions today originate on ACORD forms.
The problem is that ACORD standardized documents, not data. A document has a layout. Data has a structure. The two things are not the same, and the difference between them determines whether information can move automatically between systems or whether it requires a person to read it and re-enter it somewhere else.
An ACORD 125 in PDF format contains the same information regardless of how it was filled out, but that information is not accessible to a computer the way structured data would be. The insured name is text on a page, not a value in a defined field. The FEIN is adjacent to a label, not tagged with a type. When this form arrives at a wholesaler or carrier, extracting its contents means either having a person read it or running character recognition on a rendered page, both of which introduce errors at a rate that structured data exchange would eliminate. Manual processing in insurance consistently shows error rates near 20 percent. Automated processing brings that figure closer to 2 percent. That gap is not primarily a technology problem. It is a data format problem.
ACORD's XML standards and eForms have offered a path around this for years. The standards exist. Adoption has been slow, for reasons that are not mysterious: agency management systems would need to export structured data rather than rendered PDFs, carriers would need to receive and process that data rather than route it to intake staff, and both sides would need to agree on implementation details that involve real development work. None of those steps is prohibitively difficult individually. Together they represent a coordination problem that the market has not prioritized solving.
In the meantime, the work that structured data exchange would automate gets done by people. Those people make mistakes at a rate that is normal for people doing repetitive data entry. The resulting errors create NIGOs, NIGOs create follow-up cycles, and follow-up cycles delay placements and consume time that could be spent on underwriting. This is not a problem that any single participant created, and it is not one that any single participant is positioned to fix unilaterally.
What the industry has instead is a collection of workarounds. Intake teams at wholesalers and MGAs review submissions before they reach underwriters, catching errors before they become NIGOs. OCR-based extraction tools reduce the manual burden without eliminating it. Some carriers have built submission portals that accept structured input directly, which adds a separate system for agents to populate alongside their management software. Each of these is a rational response to an infrastructure problem that hasn't been addressed at the source.
The fax machine is gone. What it produced, a document representing data rather than data itself, is still how most commercial submissions arrive.