Skip to main content

How UXPressia uses AI

Learn how UXPressia speeds up your work with the help of AI

Tanya from UXPressia avatar
Written by Tanya from UXPressia
Updated over a week ago

UXPressia users can leverage AI-based functionality that saves time and enhances productivity. This article explains how we use Artificial Intelligence (AI) and the security measures we have in place.

AI-driven capabilities in UXPressia

Our platform incorporates AI technologies from OpenAI to provide three primary capabilities:

  • Persona profile generation: users can generate and improve persona profiles based on the prompts they provide.

  • Persona image generator: users can generate custom persona images based on their prompts.

  • Journey map generation. Users can generate journey maps based on the input they provide.
    Journey map import from images: users can upload images of journey maps and personas, which our AI interprets and converts into a digital format supported by our platform.

  • Chat with a persona: users can live chat with existing personas inside journey maps they are linked to and in the Personas tool.

  • AI assistant. Users can translate journey map and persona text inside cells, expand and shorten it, simplify the language, fix spelling and grammar, adjust the tone, add region- and industry-specific details, add examples, and create customer quotes.

  • Workspace insights. Users can generate meaningful insights based on workspace content like journey maps, personas, touchpoints, metrics, and documents. This feature requires indexing a workspace content.

AI techniques and tools used

UXPressia's AI capabilities are powered by the following techniques:

  • Large Language Model (LLM): Used to generate customer journey maps, persona profiles, and persona images based on user inputs and chat with the user on behalf of a persona, as well as translate persona content into other languages.

  • Vision features: Employed to decipher images and accurately reconstruct journey maps inside the platform.

  • Embeddings and retrieval-augmented generation (RAG): Used in Workspace insights. Selected workspace content is converted into numerical embeddings and stored securely to support fast and relevant retrieval during insight generation.

The first two tools are integrated into UXPressia through the OpenAI API, which allows us to harness the power of AI without requiring any training data from our users. For Workspace insights, we use Pinecone to create and store numerical representations (also known as embeddings) of workspace data. These embeddings are later used to generate relevant answers to user requests.

Data handling and security

Here's how we handle and protect your information:

  • Data usage. The solution operates by processing user-provided prompts (for persona profile and image generation) and images (for journey map and persona reconstruction), which do not contain personal information unless the user explicitly includes it. UXPressia does not use your data to train or enhance AI models. Our AI models are pre-trained by OpenAI and are not further trained on customer data.

  • No data storage: We do not store any customer-specific metadata or personal information. Data is processed in real time through OpenAI's API; we do not store any metadata related to these operations. No customer-specific metadata is retained for storage or future use, either.

  • Data in transit: All data, including prompts and images submitted by users, is encrypted and secured using Transport Layer Security (TLS).

  • Workspace indexing and Pinecone: When using Workspace insights, selected workspace content is converted into embeddings and stored securely with Pinecone. This trusted third-party subprocessor enables fast, accurate search as part of insight generation. Personally identifiable information (such as document owners, tag and Chunk owners) is excluded from indexing.

  • External data sources: Our AI models are pre-trained on publicly available datasets and do not use customer data for further training or improvement.

Generative AI capabilities

UXPressia integrates Generative AI capabilities directly into the platform:

  • Generative AI usage. Our platform uses Generative AI technology for creating persona profiles and images and analyzing journey maps. Generative AI is embedded into UXPressia through the integration of OpenAI's API.

  • Shared infrastructure. All AI processing occurs occurs on the infrastructure provided by OpenAI and Pinecone, ensuring security, scalability, and performance.

Workspace insights and indexing

Workspace insights are powered by indexed workspace content. Before the first generation of insights, users must index their workspace. This process converts selected data into a format optimized for AI-based retrieval and analysis.

Indexing is handled securely through our integration with Pinecone, a third-party subprocessor. No personally identifiable information is included in this process.

Limitations and responsibilities

  • No dedicated instances: We do not currently offer dedicated AI instances for individual customers, as our solution leverages the shared infrastructure provided by OpenAI and Pinecone.

  • OpenAI's role: OpenAI is responsible for continuously validating and improving the AI models we use, including testing for fairness, transparency, and security.

  • Pinecone’s role. Pinecone is a third-party subprocessor. It stores numerical representations of selected workspace content in a secure cloud infrastructure. Pinecone does not access or interpret the original content beyond what's necessary for enabling fast and relevant retrieval.

Note: To disable AI capabilities in your workspace, contact our support team.

If you have any questions or concerns about our AI capabilities, don't hesitate to contact our support team at support@uxpressia.com or via live chat.

Did this answer your question?