Unlike other AI tools, Protégé leverages our comprehensive repository of authoritative content and can access the user’s own proprietary documents, past work, and other firm or organisational knowledge. It uses this to provide a personalised experience and tailored output. Personalisation settings are user-controlled. Proprietary documents are stored securely using the RELX Responsible AI Principals and commercial grade cloud infrastructure.

AI personalisation adapts content, services, experiences, and output to match an individual's needs and preferences. Through AI personalisation, Protégé helps tax  professionals customise their interactions and AI-enabled output to meet their unique needs and preferences. Here are some examples: 
  


Document analysis and review: Protégé for Tolley+ AI can analyse a client's specific documents, such as contracts or accounting files, and highlight relevant information, potential issues, or areas of concern based on the unique details of their situation. 
  


Tax research and case law analysis: Protégé for Tolley+ AI can search through vast databases of legal precedents and statutes, identifying relevant cases and legal principles that are most pertinent to a client's specific tax matter. 
  


Personalised tax guidance: Protégé for Tolley+ AI can provide tailored tax information and explanations to clients based on their individual queries or tax needs. 
  


Tailored contract review: Protégé for Tolley+ AI can review and analyse contracts specific to a client's business or tax matter, identifying potential risks, inconsistencies, or areas that require attention based on their unique circumstances. 
  


The result is personalised responses and work product tailored to each user. Regardless of use case, the user is always in control.

The Vault is a secure repository within Tolley+ AI where professionals can upload, save, and organise key tax documents. Unlike standard uploads, documents stored in a Vault are retained securely until the user chooses to remove them, creating a personalised knowledge base. Vault capabilities include:

• Document Storage: Create up to 50 Vaults, each holding 1–500 documents.
• Secure Access: All content is encrypted and private; only you and authorised colleagues can access it.
• Smart Retrieval: Use uploaded materials as a basis for summarisation, or drafting – the AI will cite from your Vault alongside Tolley sources.
• Flexible Use: Add or remove documents at any time, with retention controlled by the user.
• Session vs Vault Uploads: Files uploaded outside of a Vault are automatically purged at the end of the session, whereas Vault-stored content is retained for ongoing use.

The Vault ensures professionals can combine Tolley’s authoritative content with their own internal documents for faster, context-rich outputs.

Extractive AI focuses on extracting specific information or insights from existing data sources, such as documents, images, or audio files. Extractive AI excels at information retrieval, data summarisation, and precision in response. It uses natural language processing (NLP) to precisely identify and extract relevant text or data from lengthy or unstructured documents or datasets based on keywords or queries. It became widely used in the 1990s and 2000s due to advancements in machine learning and the increased availability of digital text. It is applied in many familiar consumer technologies such as search engines. Extractive AI is applied in many of Tolley’s solutions to search through vast, unstructured data and return precise, clear responses. For example, in the Ask a Question of Vault functionality, Protégé in Tolley+ AI leverages extractive AI—as well as other forms of AI discussed below—to retrieve relevant information from the Vault and Tolley’s own authoritative content. 
  


Generative AI specialises in creating new content, such as text, images, music, or code, by learning from patterns in existing data. Generative AI, also known as genAI, excels at content creation, data-driven pattern recognition and application, and creative versatility. While generative AI originated through applications machine learning in the 20th century, widespread use and consumer awareness occurred more recently due to major advancements in deep learning and large language models (LLMs). It powers tools like ChatGPT, Claude, and DALL-E, which use vast quantities of underlying data to generate high-quality text, images, and other content. Tolley applies generative AI in Tolley+ AI to offer an AI assistant experience which includes features like conversational search, document summarisation, document drafting, and more. Among the company’s unique generative AI achievements are the management and minimisation of hallucinations, which are inaccurately produced creation of false or misleading content. Protégé, which is the next generation of the AI assistant experience, applies generative AI to power most of its features that involve the production of text, including but not limited to Conversational Search, Drafting or generating a Timeline. 
  


Agentic AI focuses on creating autonomous AI agents that can perform tasks, make decisions, and adapt based on real-time inputs. AI agents are software entities that can perform tasks autonomously by perceiving its environment, making decisions, and taking actions to achieve specific goals. Agentic AI excels at autonomous task execution, real-time adaptation, and workflow automation. Also, with its roots in the 20th century, agentic AI gained more widespread use in the 2010s and 2020s with the rise of machine learning, neural networks, and reinforcement learning. Agentic AI is applied in autonomous vehicles and intelligent AI assistants. Agentic AI is an area of continuous focus for Tolley and many other technology companies and will be a cornerstone of many ongoing advancements in the information services industry. 
  


It’s important to note that these forms of AI have been areas of scientific exploration for decades and already appear in technology that individuals use every day, such as Google Search, iPhones, and Microsoft Word. While individuals may not need to understand the full breadth and history of artificial intelligence for their daily work tasks, understanding the definitions and differences can be helpful to identifying assessing the right tools for their personal and professional use.