AI Agent for work

Frequently asked questions

• Which language models do we get access to?

  Ayfie Agent offers multiple models, accessible without separate subscriptions. Toggle between them easily in your chat.
  
  We offer leading AI language models like for instance ChatGPT, Mistral, Claude, and Grok out of the box. If you need a model we do not offer, reach out—we will set it up for you.
  
  Ayfie mirrors your company's policies and you can select which models your employees can access.

• Do we need a subscription for each language model?

  No, all language models are included, with consumption costs covered unless using on-prem infrastructure.

• Who owns our data?

  Your data is yours, secure within our protected Azure instances in Europe. Your data is not used for model training either. 

• Do the language models train on our data?

  No, your data is not used for model training and our technology is running on Azure instances in Europe.

  We run the language models in private mode, and only share chuncks of data that is necessary to complete the task and will not provide the LLMs with context it can train on.

• Can you deliver in our infrastructure?

  Yes, one of our core values is the ability for you to set up the complete solution within your own infrastructure.

• Is there a binding period and can we terminate or adjust the subsciption?

  The subscriptions are flexible—you can scale the number of users up or down as needed, and Ayfie does not enforce any long binding periods beyond the agreed terms. If the solution is no longer needed, the subscription can be terminated without affecting your original data or systems.
  

• Do we need to clean our data before we get started?

  In general, if you know the quality of your data you do not have to clean your data to be AI ready.

  However, in Agent, you can choose the sources you want the model to base its response to. If you know your data is not accurate, the response will be based on that data. Shit in, shit out.

  You can use Agent to clean and update your current data. We also have a separate solution for this, Ayfie Auditor.

• How do we avoid AI hallucinations?

  Ayfie Agent minimizes AI hallucinations through:

  • Secure data management: By processing data within your infrastructure, we prevent inaccuracies from external sources.
  • Customizable data inputs: You choose verified information sources, guiding AI to reliable data.
  • Continuous improvement: Feedback loops refine AI responses over time, enhancing reliability.

  Our technology is built to ensure precision and trust in every interaction.