Yes, Hevi AI can read scanned / image PDF documents as well as searchable text-based PDF documents. It uses AI-powered Optical Character Recognition to interpret text from images.
Hevi AI only accepts PDF documents, so to use a Microsoft Word (.docx) document we recommend saving or printing to PDF and then uploading that file.
Yes, Hevi is set up to support hundreds or thousands of users collaborating across multiple projects and multiple contracts.
Yes, Hevi has a User Management System which allows you to assign roles to users with different permission levels. For example, you can restrict certain users from downloading raw contract documents or from accessing certain projects.
Yes, Hevi has no lock-in contract and no minimum term so you can cancel at any time. Your service will continue until the end of the month for which you have paid.
We know that we are handling sensitive information and have therefore implemented robust controls to ensure the privacy and security of our customers' data.
Our application and database is hosted on Microsoft Azure to take advantage of its unparalleled built-in security services.
User authentication is all managed through the industry-leading external provider Auth0 and works through Single Sign-On. This means users log on through Office 365 and we do not store or manage any log in information.
Data communicated to external APIs is fully encrypted during transfer using TLS and is not stored after processing. We only use external API services from global Tier 1 technology companies that we have validated as fully compliant with all relevant information security protocols.
We also have the option to deploy within a fully enclosed environment, meaning no data is sent to external APIs, for projects that require the highest level of security.
Yes, Enterprise customers can choose to have Hevi AI deployed in a dedicated or on-premise environment so the data is never sent to any third-party servers.
No. Your data is never used to train any AI model. Hevi AI uses a pre-trained AI model which answers questions based on the context it is provided. After it answers the question, that context is forgotten.