automat: AI-Driven RPA and Document Processing

AI-driven RPA platform using generative AI and computer vision for workflow automation across Windows, Citrix, and web portals.

Overview

Founded in 2023 by former Google engineers Gautam Bose and Lucas Ochoa, automat positions itself as an AI-powered alternative to legacy RPA vendors like UiPath and Automation Anywhere. The San Francisco startup raised $15.5 million in Series A funding led by Felicis in November 2025, bringing total funding to $19.25 million with participation from Initialized Capital, Khosla Ventures, Y Combinator, K5 Global, and Input Capital.

Unlike pure-play IDP vendors like Rossum or ABBYY, automat combines document extraction with UI automation through transformer-based models that convert screen recordings into deployed workflows. The platform targets enterprises frustrated with traditional RPA's brittleness, with production deployments running 1,400 executions per day in a high-stakes banking environment.

In early 2026, automat publicly launched Document Studio, a no-code IDP product that originated as an internal tool built to unblock RPA customers stalled at the document ingestion step. The move converts a recurring implementation workaround into a standalone product line and deepens the pipeline from document extraction into automat's broader workflow automation platform.

How automat Processes Documents

automat's processing architecture converges intelligent document processing with agentic AI. The core RPA engine uses transformer-based models to mimic human behavior across Windows, Citrix, and web portals simultaneously, while a proprietary agentic layer integrates click-based UI agents, document extraction, and API connections using vision-language models and LLMs.

Document Studio, launched in early 2026, exposes this extraction capability as a standalone no-code interface. Operations teams upload sample documents, define extraction fields, train custom models against specific document layouts, and deploy via API endpoints or dashboard-triggered workflows without engineering involvement. As CEO Lucas Ochoa explained at launch: "We've repeatedly observed that customers need to extract data from various documents before they can begin further RPA automation with us. We've built tools internally to quickly solve this for customers, and today we're releasing this as an offering for everyone to use."

The platform's text-to-action approach aligns with the direction IDC's 2025-2026 MarketScape identified as the category's defining challenge: moving beyond unstructured document processing toward extracting meaningful insights and building end-to-end automation workflows powered by multimodal generative AI. Current production deployments include processing millions of documents for a multinational insurance firm, handling KYC workflows across eight systems for an international payment processor, and automating mortgage workflows including PDF classification and appraisal data extraction for AmeriTrust.

No pricing, GA status, or accuracy benchmarks for Document Studio have been disclosed. Buyers evaluating the platform cannot yet make a cost comparison against established IDP alternatives such as Hyperscience, Instabase, or AWS Textract.

Use Cases

Insurance

automat automates claims document processing across multiple legacy systems, extracting data from claim forms and PDFs, populating Citrix-based claims systems, and routing documents through approval workflows. The platform processes millions of insurance claims documents for multinational carriers. Document Studio extends this to upstream ingestion, allowing operations teams to define extraction schemas for claims forms before routing structured data into RPA workflows.

Banking and Financial Services

automat handles KYC workflows across eight different systems for international payment processors and executes mortgage workflows with PDF classification and appraisal data extraction, delivering seven-figure OpEx savings for enterprise customers. The banking deployment running 1,400 daily executions represents the platform's most cited reliability benchmark, though no independent verification has been published.

Healthcare

The platform processes patient forms and enters data into Windows-based health information systems with HIPAA compliance controls. Document Studio's target vertical list explicitly includes healthcare, positioning automat for patient record digitization and prior authorization workflows where unstructured document ingestion precedes downstream RPA steps. Unstract takes a comparable no-code LLM approach to the same document ingestion problem, offering a useful point of comparison for buyers evaluating open-source alternatives.

Mortgage

AmeriTrust is a named production customer for mortgage document automation, with automat handling PDF classification and appraisal data extraction as part of end-to-end loan origination workflows. See the mortgage document automation guide for broader context on this use case.

E-Commerce

E-commerce is named as a Document Studio target vertical, though no production deployments or customer references have been disclosed in this segment. Buyers evaluating no-code document extraction for order and invoice workflows may also want to review DOConvert, which targets supply chain and manufacturing documents with ERP integration.

Technical Specifications

Resources

• Website
• Series A Funding Announcement
• Document Studio Launch Post
• IDP Newsletter Coverage - Document Studio
• Agentic Document Processing Guide
• Data Extraction Capability Overview

Company Information

Location: San Francisco, California
Founded: 2023
Founders: Gautam Bose (former Google), Lucas Ochoa (former Google)
Total Funding: $19.25M ($3.75M seed, $15.5M Series A)
Lead Investors: Felicis (Series A), Initialized Capital (seed)
Additional Investors: Khosla Ventures, Y Combinator, K5 Global, Input Capital