KAPTO AI: Document Automation for Insurers, Industry

On This Page

Overview
KAPTO certifies the data before it acts on it
The developer hub is for partners, not self-serve signups
Use cases
Insurance claims intake
Broker statement reconciliation
Renewal risk assessment
Shipping notes and order intake in manufacturing
The homepage figures run ahead of the case studies
Technical specifications
Resources
Company information

KAPTO AI automates document-driven back-office processes for European insurers and manufacturers: it reads incoming emails and attachments, checks them against business rules, and writes the outcome into the customer's existing systems.

KAPTO homepage with the headline Up to 50% lower costs, Up to 80% faster operations

10M+Pages processed in production (vendor-stated)

~85%Straight-through processing on Poltrona Frau shipping notes

6Published customer case studies

6 weeksTypical first rollout, priced per page

Overview

KAPTO sells finished processes rather than an extraction API. Where most vendors in this IDP vendor directory lead with field accuracy, KAPTO's homepage talks about claims opened, broker statements reconciled and shipping notes posted to the ERP, and describes the product as an "industrial-grade AI digital workforce" that reads business content, validates it and "executes actions in existing systems" with no changes to core platforms. The unit of delivery is a workflow the company configures and runs with the customer, typically live within six weeks and billed per page processed.

The customer base sits in two verticals. In insurance, KAPTO publishes case studies for Helvetia France, the Italian specialty insurer REVO and an unnamed Italian insurer, and shows Net Insurance among its logos. In manufacturing, the references are the furniture makers Poltrona Frau and Cassina, both part of Haworth Lifestyle Group, and Ercos. The company is registered in London, while the directors on the UK register are Italian nationals and most of its named customers are Italian. By its own count, KAPTO runs more than 30 workflows in production for more than ten enterprise teams.

That positioning puts KAPTO against two different sets of competitors. In insurance intake and underwriting documents it meets Cytora, SortSpoke and Convr; in order and shipping-note automation it meets Workist and Klassif AI. It differs from both groups in delivering the full process, including the step where the data changes a record in the customer's system.

KAPTO certifies the data before it acts on it

KAPTO's architecture page describes five steps. It ingests from email, attachments, PDFs, scans, portals, SFTP, scanners, mobile apps and APIs. It then "acquires knowledge", interpreting content in context rather than pulling isolated fields, which covers the OCR, classification and extraction stages. In the third step, "certify knowledge", it checks what it found against business rules, structured data and other trusted sources. It then decides the next action, either by deterministic logic or AI-supported judgment, and acts: routing work, triggering steps, updating systems and notifying people.

The certify step is what separates this from a document parser. On Poltrona Frau's shipping notes, extracted lines are reconciled with order data before the ERP and warehouse system are updated; on REVO's broker statements, figures are compared with internal records and discrepancies are flagged while processing runs. Confidence scores and rule thresholds decide whether a case runs straight through or goes to human review, and every decision, validation and action is logged for audit.

KAPTO does not name its models. The company says it uses "the best available approach for each step, from deterministic logic and rules to proprietary models and LLMs", without disclosing which LLM or OCR providers sit in the pipeline. Buyers subject to the EU AI Act or DORA outsourcing rules should ask for that list, since it determines where their documents are processed.

The developer hub is for partners, not self-serve signups

A developer hub publishes an OpenAPI REST reference grouped into Administration, Process and Elements APIs, plus extension APIs for notification, post-extraction and pre-evaluation hooks. Those hooks let a partner inject custom logic before a document is evaluated and after data is extracted, the points where workflow integration usually happens. Extraction distinguishes simple fields from unbounded ones such as repeating line items, and there is a human annotation interface. The hub describes itself as serving "developer partners and customers"; there is no public trial or pricing, and access comes through a sales engagement.

Use cases

Insurance claims intake

An unnamed Italian insurer receives claims by email and by PEC, Italy's legally binding certified email, with CAI joint accident report forms, photos and supporting documents attached. Before KAPTO, staff opened each message, checked the attachments and keyed the data by hand. KAPTO now extracts the claim data, sends it to the claims system, files the documents in the archive and updates operational dashboards. The insurer reports 70% of cases processed automatically, 80% less processing time and over 95% extraction accuracy.

Broker statement reconciliation

REVO receives monthly statements from multiple brokers, each in its own template and level of detail. KAPTO extracts and normalizes the data, compares it with REVO's internal records, flags discrepancies and sends the output to REVO's management system. Staff handle only the discrepancies, and 70% of documents are processed automatically.

Renewal risk assessment

At Helvetia France, renewals across large portfolios of insured sites depended on reading fire risk assessments and other exposure documents that varied in format and detail from site to site. KAPTO structures the risk information as the documents arrive, and the case study reports 95% of the process handled automatically with 80% less processing time. "What matters most is working with someone you can trust," said Alexis Cordero, AI, IoT and tech innovation leader at Helvetia France, who also joined KAPTO co-founder Gabriel De Dominicis for a February 2026 webinar on moving AI from pilot to production.

Shipping notes and order intake in manufacturing

Poltrona Frau runs two supply chains with different rules: its semi-artisanal Residential division and Interiors in Motion, which follows automotive production cycles. KAPTO handles shipping notes from intake through validation to the ERP and warehouse management update, applies the right rules per division, and reports about 85% straight-through processing at over 99% accuracy. When Poltrona Frau changed its product, order and supplier master data in an ERP transition, the workflow was adapted in two weeks. "We've fully automated our shipping notes management, eliminating manual data entry," said Ervino Riccobon, Chief Operational Excellence at Poltrona Frau. Sister company Cassina reports similar results on shipping notes, and Ercos reports 99.5% accuracy on order intake from customer emails and PDF attachments, according to the case study listing.

The homepage figures run ahead of the case studies

KAPTO publishes more customer evidence than most vendors its size, with six case studies, two named customer quotes and a customer on camera. The headline numbers on its homepage, though, don't always match the case studies behind them. The homepage credits REVO with a "90% workforce cost reduction"; the REVO case study reports 70% of documents processed automatically and gives no cost figure. The homepage lists "80% faster claims processing" for Helvetia France; the Helvetia case study concerns renewal risk documentation, not claims. The aggregate claims of up to 50% lower costs and 80% faster operations are vendor-stated ranges. Ask for the measurement behind whichever figure a business case will rest on.

The same applies to compliance. The architecture page describes KAPTO as "GDPR, DORA and AI Act ready", with encryption in transit and at rest, logged access and a separate environment per client, deployed in the cloud or on the customer's premises. Those are design statements. The site publishes no ISO 27001 certificate, SOC 2 report or other third-party security attestation, which regulated insurers will request during vendor onboarding.

Technical specifications

Feature	Specification
Delivery model	Configured end-to-end workflows, from intake to system update
Pipeline	Ingest, acquire knowledge, certify knowledge, decide, act
Inputs	Email and shared mailboxes, PEC, attachments, PDFs, scans, portals, SFTP, scanners, mobile apps, APIs
Document types	Claims and CAI forms, broker statements, risk assessments, policy documents, pension transfers, health reimbursement claims, customer orders, shipping notes, invoices, customs data, quality and lab certificates
Validation	Business rules, structured data and trusted sources; discrepancy detection
Human-in-the-loop	Confidence and rule thresholds route exceptions to reviewers; human annotation
Outputs	Updates to claims, policy, ERP, WMS, QMS and LIMS systems through APIs and existing integration layers
Developer access	OpenAPI REST reference; Administration, Process, Elements and extension APIs with pre-evaluation, post-extraction and notification hooks
Models	Rules, proprietary models and LLMs; providers not disclosed
Deployment	SaaS with a segregated environment per client, client cloud or on premises
Compliance	"GDPR, DORA and AI Act ready" (self-declared); encryption in transit and at rest
Certifications	None published
Reported results	70-95% automation rates; 85% STP and 99%+ accuracy on shipping notes; 99.5% accuracy on order intake
Scale	10M+ pages, 1M emails a year, 30+ production workflows (vendor-stated)
Pricing	Per page processed, not published
Rollout	About six weeks for a first workflow