We won the Open Agentic AI Innovation Award at the first AAIF Technical Excellence Awards.We won the Open Agentic AI Innovation Award Read why

Intelligent document processing with AI agents

Documents read by agents and checked before use

For a US health insurer, agents that turn scans and handwriting into checked fields cut claim review from 3 hours to 8 minutes.

Book a free consultation How we build agents
document-intake

  1. Operations analystYour team

    A supplier sent a scanned contract bundle with tables and stamps. Can the intake agent extract it?

  2. Intake agentAgent

    1. parser.read_layout tables, stamps and handwriting
    2. classifier.split contract, annex, certificate
    3. model_a.extract parties, dates and amounts
    4. model_b.extract the same fields, independently

    Fields extracted. The two readings disagree on the renewal date in the annex, so it needs a check.

    Needs a personReadings disagree on one fieldRenewal date in the scanned annex
    Resolved by Operations analyst

  3. Intake agentAgent

    1. dms.write_record fields linked to source pages

    Record saved. Each field links to the page it came from, so anyone can check the source.

Example data. An agent at work in a team channel, with a person approving the step you choose.
Mixam: conversion to a final sale from about 20% to about 40%
Where agents help

Documents pile up because people retype them.

Agents read and check the documents, and people review only what is unclear.

  1. 01

    Where it stallsForms, scans and PDFs never look the same twice.

    What we do Agents parse layouts, tables and handwriting and split mixed bundles by document type.

  2. 02

    Where it stallsOCR output still needs a person on every field.

    What we do Two readings are compared in code. Only fields that disagree go to a reviewer.

  3. 03

    Where it stallsExtracted data has no link back to the page.

    What we do Each field points to its source page, so a reviewer or auditor can check it quickly.

What we build

What we build for document processing

Whatever comes in, every project is built from the same parts below, set up for your files and your systems.

Documents we process

  • Invoices
  • Purchase orders
  • Contracts
  • Claim forms
  • ID and KYC files
  • Shipping papers
  • Reports and statements
  • Emails with attachments

How we build it

Not seeing your case? Bring the files that cause you trouble and we start from those.

Tell us about your workflow
How we build agents

Agents that work as part of your team

They work in the tools your team already uses, take on whole tasks and stop for a person where you decide.

  1. Routines: agents that start on a schedule, on new GitHub issues or on an API delivery.
  2. Connected systems: GitHub, Linear, Notion, Slack, Jira and other tools an agent may call, each connected with its own sign-in.
  3. Approvals log: each action that waited for a person, with who asked, who decided and when.
Routines: agents that start on a schedule, on new GitHub issues or on an API delivery.
Connected systems: GitHub, Linear, Notion, Slack, Jira and other tools an agent may call, each connected with its own sign-in.
Approvals log: each action that waited for a person, with who asked, who decided and when.

Screens from the workspace our agents run in, with example data.

Where it runs

Teams and industries that use it

Each page shows where agents fit in that team or industry.

Engineering in the open

We help build what agentic AI runs on

Open standards, open-source tooling and published research, all public, so you can check our engineering before you hire us.

Shaping agentic AI

Agentic AI Foundation

Vstorm was the first agentic AI engineering company to join the Agentic AI Foundation and received its 2026 Open Agentic AI Innovation Award.

Read the announcement

Open source

We build open-source agent tooling that other engineering teams install

10M+ PyPI downloads
4,000+ GitHub stars
33 public repositories

Framework contributors

When the framework breaks, we fix it at the source

We maintain our own open-source agent framework and contribute fixes to the frameworks our agents run on, through official partnerships with their teams. Breaking changes reach us before they reach your codebase.

Research

Our engineers and PhDs publish peer-reviewed research

Peer-reviewed papers on agent architectures, evaluation and AI security, and an open benchmark anyone can rerun against our results.

Track record

Frontier AI teams now build on what we shipped first

Our client agents were in production before most of today's agent frameworks had a stable release. You get a team that already knows what breaks between a promising demo and a system people rely on every day.

Case studies
All case studies
Engagement models

Ways to start

Each path starts with a free call with a founder or senior AI engineer. A build goes ahead only after its Proof of Value meets the agreed threshold.

Start with one workflow

We map one workflow with your experts and agree a success threshold before any build.

Best whenYou want the success test agreed before you commit to a build.

Book a free consultation
  1. Process redesign and build Best whenYou want one team accountable for the outcome.
FAQ

Intelligent document processing, answered

What is intelligent document processing? +
Intelligent document processing (IDP) turns scans, PDFs, forms and emails into structured, checked data. Agents parse the layout, classify and split mixed bundles, extract the fields you need and validate them against your rules and systems. This is also called AI document automation. Fields the checks cannot confirm go to a person, and each value stays linked to the page it came from.
Which documents should we automate first? +
Start with a document type that arrives often, has a known set of fields and takes real reading time, such as claim files, supplier contracts or property reports. Put your hardest examples in the test set, because tables, stamps and handwriting are where parsers fail. Agree an accuracy threshold per field with the people who check those documents today.
Have you built intelligent document processing before? +
Yes. For a US health insurer we built a claims system where LlamaParse parses the file, GPT and Gemini read it in parallel and code compares the two readings. Claim processing fell from 3 hours to 8 minutes, with a specialist still deciding. For Mapline.AI we chose LlamaParse Premium after Mistral OCR missed tables in municipal PDFs, and real estate due diligence fell from weeks to minutes.
How accurate is AI document extraction, and what happens when it is unsure? +
Accuracy depends on your documents, so we measure it per field on your own files before launch. Two models read each document and code compares their answers. When they agree and the checks pass, the field is saved. When they disagree or a value breaks a rule, the document goes to a reviewer with that field marked. Corrections go back into the test set.
How do you know it works before the full build? +
We agree a success threshold with your team before the build and test the agent on your own cases in a Proof of Value. It measures accuracy and running costs, and the build goes ahead only when the threshold is met. Corrections from your team go back into the test set, so every later version is checked against them.
Who owns the system, and which models does it run on? +
You own the code, prompts, evaluation sets and rules by contract. The agents run in your environment on hosted or open-weight models. The model sits behind an interface in code, so you can change it when accuracy, speed or your hosting rules change, without rebuilding the agent.
Start with one workflow

Test one workflow on your own data

A Proof of Value shows accuracy and running costs before the full build.

Book a free consultationRead case studies
From the blog

Reading on document processing

All articles
Engineering

What is Information Extraction using LLMs?

IE emerges in making sense of this data deluge, transforming unstructured data into structured actionable information. AI for companies
Engineering

When clean text is not enough: structured extraction for RAG

When Clean Text Is Not Enough: Structured Extraction for RAG enables richer retrieval and accuracy—explore smarter solutions.
Engineering

Beyond Frontier Models: Testing Lightweight LLMs for Document Processing in RAG

Why bother with small models when capable LLMs are available? Because businesses may face stricter requirements than inidividual users