Data retention policy
Data submitted will be stored in your selected Aha! account(s). Records created within Aha! are maintained for the lifetime of the account unless they are explicitly deleted. Upon termination of service, data may be deleted from the application on request, or, if this request is not received, deleted after an inactivity period. Backup and log data are automatically deleted after one year according to the backup retention policy.
Data archiving and removal policy
Aha! provides deletion features as part of the product. This includes features to help customers comply with their own requirements such as GDPR for deletion or removal of personal information. Aha! provides an online profile center where Aha! users can review and update data and in some cases delete it (such as ideas portal users). For assistance in data deletion or removal, please contact support@aha.io.
Data center location(s)
Ireland, United States, Germany
Data hosting details
Aha! uses Amazon Web Services for data center services
Data hosting company
Amazon Web Services
App/service has sub-processors
yes
Guidelines for sub-processors
App/service uses large language models (LLM)
yes
LLM model(s) used
OpenAI GPT, Anthropic Claude (via Amazon Bedrock), Amazon Nova, Google Gemini
LLM retention settings
Aha! AI provider API calls disable storage and retention per enterprise agreements; temporary in-memory caching aids performance. Aha! saves chat history outside LLMs. Review or delete chat prompts and responses in AI assistant > History.
LLM data tenancy policy
Aha! uses AWS, Google, and OpenAI platforms designed with account-level data segmentation controls. These AI providers implement segmentation as part of their architectures. Aha! uses pre-trained models that do not include any Aha! customer data.
LLM data residency policy
Aha! stores no data in models or LLMs and offers no fine-tuning. Prompts go to AWS, Google, and OpenAI models via secure, encrypted API calls within your Aha! account's region or the USA, depending on regional model availability.