GENERAL
What types of AI does PandaDoc use?
PandaDoc leverages conversational AI, natural language processing (NLP), and generative AI technologies to operate capabilities like content summarization, agreement Q&A, and contract data extraction.
What AI models does PandaDoc use?
PandaDoc uses a combination of local AI models and select enterprise-grade third-party LLMs (e.g., OpenAI, Anthropic), chosen for quality and security. Any customer facing AI features are supported by internal AI models or vendors provided on our subprocessors list.
PandaDoc is continuously improving our AI capabilities, ensuring you get the best performance. Our AI systems are built with industry standard security, privacy, and accuracy as foundational principles.
How do I know when AI is part of my product experience?
Look for the sparkle icon to identify AI-powered features as we expand visual indicators across the product.
DATA USE & PRIVACY
Does PandaDoc use my data to train AI models?
No. Customer Content is not used to train any AI LLMs. Anonymized and aggregated AI outputs, metadata and analytics such as usage data from your sessions may be used to improve our models unless you opt out of AI features.
Does PandaDoc share my data with third-party AI providers?
To the extent that a third-party AI provider is listed as a subprocessor on our Subprocessors page, PandaDoc does share data with its subprocessors pursuant to our MSA and DPA which specifically states that PandaDoc may utilize Subprocessors to act on its behalf in connection with providing the Services provided that (a) such Subprocessors are subject to confidentiality and data security obligations that are substantially as protective as those set forth in our MSA; and (b) PandaDoc is responsible for such Subprocessors’ acts and omissions in relation to PandaDoc’s obligations under the MSA.
Are third-party AI providers allowed to retain or train on my data?
No. Third-party AI providers that are used as subprocessors by PandaDoc are prohibited from training on and retaining customer data under PandaDoc’s enterprise agreements with such providers.
Does AI processing follow my data region?
AI processing is currently only available on the US hosted PandaDoc instance, and therefore processed in the United States.
⚠️ At this time, AI features are not supported for EU-hosted accounts. If your account is provisioned in the EU, AI will be disabled by default. We’re evaluating options and will share updates as they become available.
Who owns the AI-generated content?
You own the output of the AI-generated content. It’s your responsibility to review AI outputs for accuracy and legal sufficiency before use.
SECURITY & COMPLIANCE
How does PandaDoc keep my data secure?
PandaDoc’s industry standards security practices can be viewed on our website. We keep your data safe through best-practices like:
Data is encrypted when traveling and at rest
All actions follow your role-based permissions. AI can only access data within your workspace that you can access.
AI conversations are stored in your workspace for your reference and only you can access them.
PandaDoc is SOC 2 Type II compliant with annual independent audits
Can admins disable or restrict AI features?
Yes. AI features can be enabled or disabled for a workspace.
What is the AI Addendum and does it apply to my account?
PandaDoc’s AI Addendum governs all customers’ use of AI features. It covers ownership of AI outputs, training defaults, opt-out rights, accuracy disclaimers, and third-party security obligations. You can review the full AI Addendum here.
Can I opt out of AI features for my organization?
Yes, you can opt out of certain AI functionality at any time. Opting out is go-forward only — it stops new AI processing but does not delete anonymized data that has already been used for training.
What happens to my data if I cancel my contract or opt out of AI?
Opting out or terminating your contract stops future AI processing.
Does PandaDoc guarantee the accuracy of AI outputs?
No. PandaDoc does not provide warranties for AI outputs. You are responsible for reviewing AI outputs before use.