Affinda is an Australian AI company specializing in Intelligent Document Processing (IDP). While many legacy OCR tools rely on rigid templates (breaking if a document layout changes slightly), Affinda uses advanced deep learning and Natural Language Processing (NLP) to “read” documents like a human. Originally gaining fame for building one of the world’s most accurate Resume Parsers, they have expanded into a broader AI platform that extracts structured data from invoices, job descriptions, ID cards, and government forms. They are known for a “developer-first” approach, offering robust APIs that allow software teams to embed document reading capabilities directly into their own applications (e.g., ATS, CRM, or ERP systems).
Read also: Top Australian AI startups to watch in 2026
Core Technology: Adaptive IDP
The “Vesparum” Engine: The company was incubated out of Vesparum Capital, leveraging proprietary research to build models that don’t just recognize text (OCR) but understand context.
Model Memory (RAG-style Learning): Unlike static models, Affinda’s “Agentic AI” approach learns from user corrections. If a human validator fixes a mistake (e.g., identifying a new invoice format), the model adapts, solving the “cold start” problem for custom document types.
Specialized Pre-Trained Models:
- Resume/CV Parser: Extracts 100+ fields (skills, education, work history) across 50+ languages, handling complex layouts and columns that confuse standard AI.
- Invoice & Receipt Parser: Automates Accounts Payable by extracting line items, tax codes, and vendor details.
- Search & Match: A semantic search engine that matches candidates to job descriptions based on skills and intent, not just keyword matching.
Company Profile
Founders: Tim Toner (CEO, ex-Investment Banker) and Dr. Ben Toner (Chief Scientist, PhD in Physics from Caltech).
Headquarters: Melbourne, Australia.
Group Structure: Part of the Affinda Group, which also owns Draftable (legal document comparison) and recently acquired Pathfindr (AI consultancy) for $15M (late 2025).
Funding: Raised approximately **$23M AUD** ($15M+ USD) from high-net-worth investors and family offices.
Key Customers: SEEK, Deloitte, Wattpad, Korn Ferry, Northline.
Key Use Cases
Use Case
- HR Tech / ATS: Used by recruitment platforms to parse millions of resumes instantly. It populates candidate profiles automatically, removing the need for applicants to manually re-type their CVs.
- Accounts Payable: Automates invoice processing for finance teams. It reads PDF invoices attached to emails, validates the data against the ERP, and flags only the exceptions for human review.
- Redaction (DEI): Automatically removes bias-inducing fields (Name, Gender, University, Address) from resumes before a hiring manager sees them, supporting Diversity, Equity, and Inclusion initiatives.
Why It Matters
It is estimated that 80% of enterprise data is unstructured (locked in PDFs, emails, and images). Companies spend billions manually keying this data into systems. Affinda enables the “Autonomous Enterprise” by turning this messy, human-readable content into clean, machine-actionable code, effectively acting as the “reading comprehension” layer for business software.
