
Key Takeaways
- Unstract is best for commercial submission intake (ACORD forms, loss runs and SOVs), with one output format that covers every insurer’s loss-run layout.
- Sensible suits for Rule-based extraction configurations.
- Docsumo suits Business teams building workflows.
- Nanonets suit cross-department workflows.
- ACORD forms, loss runs and SOVs fail in different ways, so test each type separately before you buy.
- Any tool feeding underwriting decisions needs a documented way to test and review its output, in line with the NIST risk framework.
For teams reading ACORD forms, loss runs and SOVs every day, Unstract is the insurance document processing tool that handles all three without templates. One fixed output format covers every insurer’s loss-run layout, ACORD checkboxes are read correctly, and Excel SOVs go in directly. Sensible, Docsumo and Nanonets each take a different route to the same documents.
ACORD lists submissions, loss runs and SOVs among the documents its own extraction tool is built to read. These three paper types decide how fast an underwriter can quote a commercial risk.
ACORD Forms, Loss Runs and SOVs in Plain Words
Each of these three documents breaks software in its own way. Knowing the difference helps you test fairly.
ACORD Forms
ACORD forms are the standard application pages used across the US, such as the 125, 126 and 140. They are dense with checkboxes and small boxes of text. A tool that reads a ticked box as an empty one changes the meaning of an application.
Layouts also change between form versions. A tool tied to one fixed layout needs fresh setup each time.
Loss Runs
A loss run is a report of a business’s past insurance claims. Every insurer formats its own, so the same facts appear under different headings, in different column orders. Some arrive as scans.
Template-based tools struggle with loss runs because a new insurer means a new layout.
SOVs
An SOV is a statement of values, usually a spreadsheet listing a business’s buildings and what each is worth. Rows run long and column names vary from broker to broker. The useful work is mapping that mess into a standard shape and comparing it with the ACORD form.
How the Tools Were Tested
Five criteria shaped the comparison, drawn from the failures above.
- New layouts: Does a first-time loss-run format need new setup?
- Checkboxes: Are ticked and empty boxes read correctly on ACORD pages?
- Spreadsheets: Does an SOV go in as an Excel file?
- Ease of setup: Who has to build and maintain the configuration?
- Cost: Is pricing public, and how does it scale with volume?
Quick Comparison: Four Tools at a Glance
| Tool | Best For | Setup Style | Pricing Model |
| Unstract | Commercial submission intake (ACORD forms, loss runs and SOVs | Plain-English, business-team setup | Free plan of 100 pages a day |
| Sensible | Rule-based extraction configurations | Configuration through API and SDKs | Public tiers, billed per document |
| Docsumo | Business teams building workflows | Workflow builder without developers | Free plan, then custom quote |
| Nanonets | Cross-department workflows | Agent-based workflow builder | Usage-based credits |
Four Tools for Reading Submission Documents
1. Unstract – Best for commercial submission intake (ACORD forms, loss runs and SOVs)
Unstract works without templates. Document Studio, with Prompt Studio, pulls the fields you want into one fixed format no matter how different the documents look. A loss run from a new insurer needs no extra setup and no template.
The reading layer is LLMWhisperer, which keeps tables and columns intact and reads ACORD checkboxes correctly. It also reads 300+ printed languages and handwriting in 12 languages, phone photos and crooked scans. SOV spreadsheets in Excel format are accepted directly, among 20+ file types that include PDF, Word and images.
Prompt Studio defines the fields to extract, and Flow builds the wider process from a plain-English description. LLMChallenge puts each field through a second model, and any mismatch lands in review. Highlight links every figure to its page.
Moody’s is among the 600+ paying customers that use Unstract. The free plan covers 100 pages a day, which is enough to test your own loss runs.
What Stands Out
- One output format covers every insurer’s loss-run layout, with no templates to build.
- ACORD checkboxes are read correctly, and Excel SOVs are taken in directly.
- The free plan of 100 pages a day lets a team test on real documents first.
- Customers can bring their own AI model and run Unstract in the cloud, on their own servers or as open source.
Points to Consider
- A pilot on real loss runs and SOVs shows a team how Unstract maps its fields.
- A bring-your-own AI model means someone picks the model and tracks its usage.
Who It Suits
Underwriting and operations teams that handle mixed ACORD, loss-run and SOV bundles and want a business team to own the setup.
2. Sensible – best for rule-based extraction configurations
Sensible is a hybrid document tool that pairs AI extraction with deterministic rules. It works through a REST API and SDKs. Its extraction configurations include loss runs, declarations pages, certificates and policies.
What Stands Out
- Validation checks, such as paid plus reserve equaling incurred on a loss run, can send a document to human review.
- Public tiers start at $499 a month on monthly billing for 750 documents, with a 14-day free trial.
- Billing is per extracted document, and the platform lists SOC 2 Type II and GDPR on every public tier, with HIPAA via a signed BAA on Enterprise.
Points to Consider
- Which ACORD forms and SOV layouts already have maintained configurations?
- Who writes and updates the configuration when a new carrier’s loss run arrives?
Who It Suits
Teams that want to shape extraction rules themselves and connect the output through an API.
3. Docsumo – best for business teams building workflows
Docsumo is a document platform with Document AI, case management and workflows. Business teams can build workflows without writing code. Its insurance coverage names ACORD 25, 125, 126 and 140, certificates of insurance, loss runs, declarations and endorsements.
What Stands Out
- A confidence-threshold review queue shows the source line for each flagged value.
- Failed checks are routed with a reason, and an audit log can be exported.
- A free plan covers 14 days and up to 1,000 pages, and larger plans are quoted by volume.
Points to Consider
- How are SOV spreadsheets mapped into its output?
- What happens when a new insurer’s loss-run layout arrives?
Who It Suits
Business teams across lending, insurance and other sectors that want a no-developer workflow builder with a review queue.
4. Nanonets – best for cross-department workflows.
Nanonets turns business processes into AI agents, with agentic data extraction at the core. Its use cases span accounts payable, order management, logistics and a first-notice-of-loss claims agent. Enterprise customers can choose VPC, single-tenant cloud or on-premises deployment.
What Stands Out
- Human approval gates route edge cases to Slack, Teams or email for approve, edit or reject decisions.
- Usage pricing runs per block, from $0.02 for simple steps to $0.30 for complex AI steps.
- A free Starter tier includes $50 in credits, and Enterprise adds SSO and data residency options.
Points to Consider
- Does it provide tested setups for ACORD forms, loss runs and SOVs, or would your team tune them?
- How much configuration does a new insurance document type need?
Who It Suits
Enterprises that want one agent platform across finance, logistics and claims workflows.
Five Checks Before You Pick an Insurance Document Processing Tool
A good demo hides most of these problems, so run each check yourself.
- Use a loss run the vendor has never seen. Ask for a layout from an insurer outside the demo set and see whether setup is needed.
- Compare checkboxes against the page. Open an ACORD form and confirm every ticked box matches the output.
- Feed in the raw Excel SOV. Skip any pre-cleaning, and note whether the columns land in a standard shape.
- Ask who owns upkeep and governance. NAIC members approved a model bulletin in December 2023 that expects insurers to keep a written AI program scaled to risk.
- Price it at your volume. Count AI usage and review time, not only the per-page figure.
Common Questions Answered
Which insurance document processing tool reads ACORD forms, loss runs and SOVs best?
Unstract is best for commercial submission intake (ACORD forms, loss runs and SOVs). One output format covers every insurer’s loss-run layout, ACORD checkboxes are read correctly, and SOV spreadsheets go in as Excel files.
Why is SOV extraction harder than ACORD form extraction?
ACORD forms follow a known structure, while SOVs are spreadsheets with long rows and column names that change by broker. The tool has to map those columns into a standard shape and compare values with the application.
How should insurers judge accuracy on loss runs?
Test on layouts from several insurers, including scans, and compare the output with the source pages. Prefer tools that check values a second time and show where each figure came from.
Does Unstract need a template for each insurer’s loss run?
No. Document Studio pulls fields into one fixed format, so a new layout needs no template. A pilot on a few real loss runs confirms this.
What documents make up a commercial insurance submission?
A submission bundles ACORD application forms, loss runs and a statement of values, along with supplemental questionnaires and the broker’s cover email.
The Bottom Line
Unstract is the insurance document processing tool we would choose for ACORD forms, loss runs and SOVs, because one format covers every layout and no templates are needed. Sensible fits teams with developers, Docsumo fits business teams that want a workflow builder, and Nanonets fits enterprises running agents across many processes.
Pick who will own the setup next year, then test each tool on a new loss run, a ticked ACORD form and a raw SOV. The tool that handles all three without rework is the one worth buying.

Peyman Khosravani is a global blockchain and digital transformation expert with a passion for marketing, futuristic ideas, analytics insights, startup businesses, and effective communications. He has extensive experience in blockchain and DeFi projects and is committed to using technology to bring justice and fairness to society and promote freedom. Peyman has worked with international organizations to improve digital transformation strategies and data-gathering strategies that help identify customer touchpoints and sources of data that tell the story of what is happening. With his expertise in blockchain, digital transformation, marketing, analytics insights, startup businesses, and effective communications, Peyman is dedicated to helping businesses succeed in the digital age. He believes that technology can be used as a tool for positive change in the world.
