Introducing GenerativeShield: integrating business data with Generative AI in a simple, secure, and versatile way

Milan, July 4th 2023 – S2E announces GenerativeShield, the Artificial Intelligence SaaS solution designed to mitigate the regulatory and privacy risks associated with the adoption of Generative AI tools in Italian and European companies.

In response to the business’ needs, GenerativeShield provides the ability to securely and easily integrate Generative AI technologies in order to enhance and develop intelligent conversational assistants that are capable of analyzing business data and providing accurate and consistent answers to users through self-service development.

S2E has chosen Amazon Web Services- which it is an Advanced Partner – Cloud regions, based in Europe, as its development and processing platform. In addition to being compliant with European GDPR regulations, it facilitates the creation of conversational modules that can be seamlessly integrated into corporate applications or used on their own.

Organizations have increasingly adopted generative artificial intelligence systems. Over time, companies became increasingly aware of the importance of integrating the technology into their organizations in a concrete way, overcoming the trade-off between the adoption of the technology in specific and private use cases involving the use of corporate data, as well as the uncertain security and confidentiality of the corporate data due to the adoption of public models such as ChatGPT.

“Generative Shield is a SaaS solution designed to mitigate the risks and complexities associated with organisations’ use of generative artificial intelligence. It became clear that the benefits of its deployment cannot be separated from a secure integration with the company’s information assets, which is necessary to provide specific answers for the different conversational use cases of companies,” says Andrea Cappelletti, Digital Transformation & Hyperautomation Business Unit Director at S2E.

“Moreover, the worth of technology becomes evident when it seamlessly merges with the practical aspects of a business and aligns with the existing applications employed by the company. That’s why our emphasis lies on a flexible solution that can easily integrate with various systems. From a technological perspective, advancements occur rapidly, parallel to the evolution of different technologies and the maturity of the market. As a result, it is uncertain which player will emerge victorious in the near future, as the market’s direction of progress remains unpredictable”, continues Cappelletti “ours is a plural approach. Our R&D is working to optimize and perfect the extraction of punctual and referenced information from corporate data, ensuring that content is only produced if the corporate data is certain and consistent with the question asked by the user, reducing the likelihood of hallucinations associated with Generative AI technologies. GenerativeShield presents itself as a platform whose operation is model agnostic, focused on the processing and extraction of certain business information, in total security and confidentiality.”.

GenerativeShield datasheet

• Early Access, for companies interested in evaluating Generative Shield’s adoption – S2E grants companies that desire to try the product free priority access to the platform for a period of 30 days following activation, up to 5 users per company. You can book HERE
• Create Private Intents integrated with business data – Generative Shield empowers you to design, set up, and deploy Private Intents that are tailored specifically for your conversational use case, enabling you to create AI-powered assistants tailored to your business needs. Explore the playground of settings and configuration, define the personality of your assistant and get ready for production with just a few clicks.
• Upload and combine your corporate Knowledge Base with your Private Intents – Generative Shield provides an intuitive, enterprise-quality environment in which you can effortlessly upload your Knowledge Base, which serves as the foundation for feeding your Private Intents. By combining Knowledge Base and Private Intents, your assistants will be able to provide accurate and documented answers to complex questions.
• Multiple Generative Artificial Intelligence Model Integration for Public Intentions – Native integration of all major public generative artificial intelligence models, including OpenAI GPT3.5 / GPT4, Aleph Alpha, Ai21 Labs, and Cohere, allowing users to select the model that best suits their generic use case.
• Mitigation of the risk of disclosure of sensitive data for interaction with Public Intentions – Reduction of the risk of disclosure of sensitive data through dedicated ‘filters’ that identify and notify users of the presence of information within prompts. Companies can then control how this data should be blocked or decide, conversely, whether to allow users to approve its external sharing.
• Integration of the services offered with existing business applications – Each of the services offered can be accessed either directly through the platform, or through widgets and/or APIs that can be integrated with business applications to enhance their capabilities.
• Preservation of user data confidentiality – ‘No-log’ policies are strictly enforced on user prompts, model responses and data used to maintain the confidentiality of the user’s data. The infrastructure is entirely based on Amazon Web Services in Europe and is GDPR-compliant with regard to the processing of personal data.
• Easy and intuitive ‘chat-like’ user experience – The user experience of the platform is like chat on desktop and mobile devices, regardless of the generative artificial intelligence models used. Users can interact with a single model or a list of public models, compare responses and decide which models to interact with.
• Native prompt engineering and optimization – Prompt Engineering is automated based on the intended use cases, both public and private, to provide users with suggestions and guidelines to optimize the prompt and achieve the desired results. Private Intents allow users to create their own prompt template based on the use case they have chosen.
• External service providers – External providers of generative AI models for public intentions, including those located outside Europe and/or not covered by the GDPR, are subject to their own privacy policies and terms of service. Platform provides a single point of control and management for all external providers of generative artificial intelligence.