Stories, guides and straight answers.
Learn how DexAI works, how teams use it and how the AI underneath is built.
Everything in one place
Case Studies
How teams put DexAI to work.
Customer Stories
In our customers' words.
Blog
Notes on financial automation.
Documentation
Guides to every part of DexAI.
AI Insights
How the AI actually works.
FAQs
Short answers to common questions.
Why DexAI
What changes when you automate.
Security & Trust
How financial records are handled.
How the AI actually works.
Short explainers on OCR and vision-language models, agents, confidence scores and matching.
- OCR / VLM
Why OCR alone isn't enough for financial documents
Reading text is the easy part. Knowing which number is the total, which line is VAT and which block is the supplier takes layout understanding.
- Agentic AI
What an agent does in DexAI
Each agent owns one job, from classification to reconciliation, and hands off to the next. Here is how the chain fits together.
- Validation
Confidence scores and human review
Every extracted field carries a confidence score. Low scores are flagged for a person, not silently posted.
- Reconciliation
How bank matching finds the right document
Amount, supplier and date are weighed together; exceptions are listed rather than guessed.
- Training
Training models on financial paperwork
Specialising on receipts, invoices and statements, and measuring every release before it ships.
- Infrastructure
Running document AI on CPU or GPU
Why DexAI's models are optimised to run on either, and what that means for deployment.
Ready to automate your finances?
Spend less time processing documents and more time running your business.