Introduction
This article sets out what each of Planhat's AI features is intended for, and the uses they are not intended for. It is part of Planhat's Documentation for the purposes of the EU AI Act, our Terms of Service and the AI Compliance Whitepaper.
Roles under the EU AI Act
As between Planhat and you, Planhat is the provider and you are the deployer of Planhat's AI features. Each party is responsible for the obligations that apply to it in that role.
What each feature is intended for
Pi (including Pi jobs). An AI agent and chat assistant for commercial and customer-management work on your Planhat data. Pi answers questions, finds information, drafts content, creates and updates records, builds AI Automations and runs scheduled Pi jobs, always within the permissions of the user it acts for. Session search is intended to help users find decisions, commitments and context. It is not intended for monitoring or evaluating the performance or behaviour of individual employees.
AI Automations. Rule-based Automations with one or more AI steps that carry out recurring commercial and customer-management tasks, such as updating records, preparing content and routing work, within the configuration and approval steps you set.
Conversation Sentiment & Classification. Helps you understand and prioritize incoming customer conversations (emails, chats and tickets) by assigning a sentiment (positive, neutral or negative) and a category. It analyses the text of the conversation and does not use biometric data. It is not intended for evaluating or monitoring the people who write or answer those conversations, including your own staff.
Conversation Summaries. Summarizes email threads, chats and tickets so that users can review them faster.
Signals. Draws your attention to events in your Planhat data that may need follow-up, by flagging them as important or as a priority.
Engagement Scoring. Shows how actively the end users of your customers engage with and adopt your product, so that you can understand the health of your customer relationships. It is not intended for decisions about individuals.
What the features are not intended for
No Planhat AI feature is designed or intended for a practice prohibited under the EU AI Act. Planhat specifies that its AI features are not to be changed into a high-risk AI system. In particular, the features are not intended to be used:
in relation to your personnel or job candidates:
to evaluate, score, rank, compare or monitor them, or their performance or behavior;
to allocate tasks based on individual behavior or personal traits;
to make or materially inform decisions on recruitment, terms of employment, promotion or termination;
for any other use listed as high-risk in the EU AI Act, such as assessing the creditworthiness of individuals;
to make decisions that have legal or similarly significant effects on a person without the human involvement the law requires, for example under Article 22 GDPR.
No Planhat AI feature is currently intended for a high-risk use. If that changes, this article will say so for the feature concerned, and that use will require additional terms agreed in writing.
If you use or configure a feature outside its intended purpose so that it becomes a high-risk AI system, you become the provider of that system under the EU AI Act and take on the obligations of that role. Our Terms of Service do not permit use outside a feature's intended purpose.
Your responsibilities as deployer
Your responsibilities include:
making sure staff who use AI features have sufficient AI literacy;
applying human oversight suited to how you use the features, including reviewing output before relying on it;
informing people that they are interacting with AI or receiving AI-generated content where the law requires it;
keeping any identification of AI involvement that the Services show, and not presenting AI output as written or approved by a person when it was not.
Customer-connected models
These statements cover the AI features Planhat provides. If you connect your own AI model, they do not extend to that model.
Related
AI Compliance Whitepaper (Section 4.1), which can be accessed via the Trust Center
Terms of Service (Section 4.8)
Last updated: October 13, 2026