AI vision parses loss runs (19 fields per claim) and equipment schedules (28 fields per unit) from PDFs, Excel, CSV, and images up to 20MB. Results saved to the client record. 10 free extraction credits included.
Documents Parsed
1,847
+124 this week
Data Points Extracted
52.3K
+3.2K
Avg Parse Time
8s
↓ 2s
Accuracy Rate
96.4%
+0.8%
What it is
A document intelligence workflow powered by advanced AI vision. Upload a loss run or equipment schedule through the Process Document slide-over, select the client from the dropdown, and the AI extracts structured data in seconds. Each parse costs one extraction credit — 10 are included free.
Insurance agencies receive documents in every format imaginable: multi-page PDF loss runs from carriers, Excel equipment schedules from insureds, scanned ACORD applications, and CSV driver lists. The AI reads these files, identifies the relevant fields, and outputs clean, structured data up to 20MB per file. A 30-page loss run becomes a searchable claims table in seconds.
Beyond document parsing, the built-in Insurstein AI (✨ sparkle button, bottom-right) handles natural-language questions about your book of business. Ask it to create a lead, move a pipeline stage, draft an underwriter email, or summarize claims history. Before executing any action, it shows a confirmation card — nothing changes without your approval. 50 free AI messages included; additional messages billed at $0.09 each monthly.
Core capabilities
PDF, Excel, CSV, JPG, PNG, WebP, HEIC — drag into the Process Document slide-over, select a client, parse.
Open the Process Document slide-over from the (+) menu and choose Loss Run Analysis or Vehicle & Equipment Schedule. Drag a file or click to browse — PDF, DOC, DOCX, XLS, XLSX, CSV, JPG, PNG, WebP, HEIC up to 20MB. AI reads the document and returns structured data. Batch uploads are supported — drag multiple files and they process sequentially.
Claim #, date of loss, claimant, type, status, paid, ALAE, reserve, incurred, subrogation — all extracted.
Loss run results appear in the client's Loss Runs tab as an expandable table: Claim #, Date of Loss, Claimant, Type, Status, Paid, ALAE, Reserve, Incurred. A totals row summarizes Paid, ALAE, Reserve, Incurred, Subrogation, Recovery, and Deductible. Results are ready for remarketing submissions immediately after parsing.
VIN, year, make, model, type, value, deductible, ownership, driver, fuel type — extracted per unit.
Equipment list results appear in the client's detail page as an expandable table showing Year, Make, Model, VIN, Type, and Value per row. Parsed schedules eliminate hours of manual data entry for large fleets and feed directly into quoting and submission workflows.
Select the client from the dropdown in the slide-over — parsed data links to their profile and Loss Runs tab.
After uploading, select the client from the dropdown (not free text — the dropdown is required for proper linking). The parsed extraction is then associated with that client and appears on their detail page. If you type a name instead of selecting from the dropdown, the document saves to Recent Extractions but is not linked to any client profile.
Natural-language AI in the bottom-right corner. Attach documents, ask questions, create records — with confirmation before any action.
Click the ✨ sparkle button (bottom-right, available throughout the app) to open the Insurstein AI chat. Attach a loss run or equipment list via the paperclip icon — Claude analyzes the file and asks what to do with the data. Ask questions about your pipeline, request a lead creation, move a deal stage, or draft an underwriter email. Before executing any action, the AI shows a confirmation card — nothing changes without your approval.
10 free extraction credits included. Credit packs from $10. Organize results in color-coded folders.
Each AI-parsed document costs one extraction credit. Plans include 10 free credits (one-time). Additional packs: $10 for 1, $85 for 10, $240 for 30, $375 for 50, $700 for 100. All parsed extractions appear in the Recent Extractions table with sortable columns (Date, Type, Client, Records, Total Value). Organize them into named, color-coded folders with private or org-wide sharing. Bulk-select rows to export CSV, move to a folder, or delete.
What they don't do
Retyping 19 fields per claim from a PDF loss run into a spreadsheet takes 20–45 minutes per report. AI vision does it in seconds.
Sending documents to offshore processing teams adds days of turnaround and creates data security exposure. Extraction credits are consumed instantly in-app.
Manually transferring 28 fields per unit — VINs, values, drivers — from insured-provided spreadsheets into your AMS is error-prone at scale. Equipment Schedule Parsing handles the whole file.
General transcription tools don't understand loss run structure or insurance fields. The AI here knows exactly what to extract from carrier documents.
When loss runs are stored as PDFs with no extracted data, finding a specific claim means opening every file. Parsed extractions are organized in searchable, sortable folders.
General-purpose AI has no access to your agency data and can't create records. The Insurstein AI (✨) reads your live book of business and executes actions with a confirmation step.
What it takes off your plate
Sends the uploaded file to the AI which reads the document layout and extracts structured fields — no OCR pre-processing or template mapping required.
Deducts one credit per parsed document, shows remaining balance in the UI, and surfaces credit pack options before blocking — with auto-pay modes available to avoid interruptions.
Adds every parsed result to the Recent Extractions table immediately, with sortable columns for Date, Type, Client, Records, and Total Value.
When a client is selected from the dropdown, links the extraction to their profile and surfaces it under the Loss Runs or Equipment Schedule tab on their detail page.
Organizes extractions into color-coded named folders with private or org-wide visibility. Bulk-select rows to export CSV, move to a folder, or delete.
Tracks AI chat messages against the 50 free message allowance, then invoices overage at $0.09/message monthly via Stripe — messages are never blocked mid-conversation.
A day in the life
8:00 AM
Three loss run PDFs arrived overnight. Open the (+) menu → Loss Run Analysis, drag all three files into the slide-over, select the client from the dropdown for each. AI processes them in sequence.
8:08 AM
67 claims extracted across the three reports. Each appears in the client's Loss Runs tab — expandable rows showing Claim #, Date of Loss, Claimant, Paid, ALAE, Reserve, and Incurred. 3 extraction credits used, 7 remaining.
9:30 AM
A prospect sends an Excel fleet list. Upload via Equipment Schedule → select the lead from the dropdown. 47 rows returned: Year, Make, Model, VIN, Type, Value per unit. Ready for quoting without re-keying.
11:00 AM
Click the sparkle button and ask: 'Summarize the loss history for Alpine Transport across both parsed reports.' The AI reads the extracted data and returns a dollar-amount summary. Confirm before anything is saved.
2:00 PM
Move the three loss run extractions into a 'Q3 Remarketing' folder (org-wide, green). Use the structured claims data and equipment table to compile the submission — minutes instead of an hour.
Daily impact
Documents Parsed
Data Points Extracted
Manual Entry Saved
Client Profiles Updated
AI vision extracts 19 fields per claim and 28 fields per unit. 10 free extraction credits included. Organize results in folders, link to client records, and feed data straight into submissions.