An Authentic AI for Contracts
Evisort AI Labs researches Natural Language Processing (NLP) and computer vision models, and develops algorithms using the most advanced forms of artificial intelligence possible, to automate search and data collection so you can find what you need in all your contracts, at any time.
Evisort AI Research
The AI research conducted at Evisort has been the subject of highly regarded research papers, including a case study by Columbia Business School.
Why AI?
Contracts are the core of any business and in today’s COVID-19 world knowing what’s in your contracts fast is paramount. Artificial Intelligence (AI) and Machine Learning (ML) can accelerate the work, reduce the need for outsourcing additional teams, and help to eliminate human error.
If you’re not familiar with AI and ML terminology let’s review them here.
Amine Anoun Head of Evisort AI Labs
Amine Anoun is the co-founder and CTO of Evisort. Prior, he served as a data scientist at Uber and wrote the walking algorithm for Uber Express Pool. He started his career studying risk management and supply chain for the FDA. Amine studied operations research at MIT.
He’s also a Forbes Technology Council contributing author!
Evisort Uses Actual AI, Not
a Rule-Based System
Out-of-the-box Evisort AI can read over 230 contract types, including blurry scans, PDFs and third-party paper, and extract over 50 clauses and metadata ready for analysis. And, it’s simple enough for you to train your own provisions.
Out-of-the-box
Evisort AI reads
over 0
contract Types
Extract over 0
clauses and metadata
ready for analysis
How do We do It?
An AI platform built through deep learning, by feeding it millions of legal contract data and allowing it to attach significance to patterns of terms and clauses leading to predicting this information on its own.
More than just OCR, Evisort trained a proprietary deep learning model to perform contextual spelling and grammar checking bringing down the word-level error rate on poorly scanned documents significantly. Evisort is also currently working on integrating Microsoft Azure Read API recent advances with our OCR capabilities to improve spelling, support handwriting, and support certain foreign languages.
Enhanced computer vision and image processing can detect objects in contracts and preserve them such as company logos, signatures (handwritten or electronic), and tables.
Tables are where key data reside in contracts and in the industry are a point of frustration, until now. Evisort accomplished the ability to read tables by running object detection models to locate table boundaries and identify individual cells, including borderless tables. Once extracted, Evisort can parse the value relationships among cells in a table, to understand the information, and expose it to the user by rebuilding tables in Word documents.