Retab has emerged from stealth mode with $3.5 million in pre-seed funding to tackle the persistent challenges developers face when applying artificial intelligence to document processing. The Paris and San Francisco-based startup addresses the gap between promising AI demonstrations and reliable production systems.
The funding round attracted early-stage investors including VentureFriends, Kima Ventures, and K5 Global, alongside notable individual backers such as Eric Schmidt through StemAI, Datadog CEO Olivier Pomel, and Dataiku CEO Florian Douetteau.
From Demo Magic to Production Reality
Founded by engineers Louis de Benoist, Sacha Ichbiah, and Victor Plaisance, Retab developed from their firsthand experience building internal automation tools for logistics workflows. The trio recognized a fundamental problem plaguing AI implementation in document processing.
“People keep building demos that look like magic, but break the moment you put them into production. We lived that pain ourselves” ~ Louis de Benoist, co-founder.
The platform serves as an AI agent that constructs document extraction pipelines through a developer-focused interface and software development kit. Users can convert unstructured documents, including PDFs and handwritten scans, into structured data by describing their requirements and uploading documents.
Orchestration Over Output
Retab’s approach centres on orchestration rather than creating new models. The platform connects with large language models from OpenAI, Google, and Anthropic, functioning as middleware between developers and model infrastructure.
The system employs model-agnostic routing, automatically benchmarking available models and assigning tasks based on specific performance criteria such as cost, speed, or accuracy requirements. This automated selection process updates continuously as superior models become available.
The platform incorporates guided reasoning and a k-LLM consensus mechanism, requiring models to follow systematic logic whilst comparing outputs from multiple models to evaluate uncertainty levels. This multi-layered approach aims to ensure consistent results across diverse document types.
“Retab is the OS for reliably extracting structured data. It wraps the best models in a layer of logic that actually makes them usable with error handling and structured outputs” ~ Louis de Benoist, co-founder.
Real-World Applications
Companies across logistics, finance, and healthcare sectors have adopted Retab to automate complex document workflows. A trucking company utilized the platform to achieve high accuracy requirements whilst reducing operational costs. Financial firms now extract insights from lengthy reports in significantly reduced timeframes.
Additional use cases span claims processing, medical records management, and identity verification procedures. The platform handles workflows involving contracts, invoices, and compliance documents, targeting the reduction of manual processes.
Infrastructure Investment
The fresh capital will support platform development, community expansion, and infrastructure scaling to accommodate growing demand from vertical AI startups and internal innovation teams. With a current team of 10 and an expanding developer community, Retab aims to establish itself within the AI infrastructure stack.
Investor Florian Douetteau highlighted the broader implications of document automation capabilities “The AI-fication of the economy depends on the capability to convert operations based on millions of documents into verified, structured data that autonomous systems can utilize” ~ Florian Douetteau, CEO of Dataiku.
Expanding Capabilities
Retab plans to extend its platform beyond traditional documents to include data extraction from websites. The company is introducing integrations with automation tools including n8n, Zapier, and Dify.
The startup envisions becoming a middleware layer connecting unstructured data with AI agents, enabling applications across loan files, contracts, and customs records. This expansion reflects the company’s ambition to support users in building and scaling data-driven workflows.
As AI adoption accelerates across industries, Retab’s focus on production reliability rather than demonstration appeal may prove decisive in determining which document automation solutions achieve widespread enterprise adoption.