Skip to main content

How AI is used in Customer Support

Josie avatar
Written by Josie
Updated this week

To improve the speed and consistency of our customer support, Famly uses a set of AI features built into our support tool, Intercom.

These tools help us answer your queries faster, surface trends in support conversations, improve the quality of our support, including our Help Centre, and provide us with valuable insights on where improvements of the Famly platform may be needed. They work by reading messages sent via our support chat and support inboxes, which means any personal data included in those messages may be processed by third-party AI systems.

Previously, only messages sent to our AI chatbot (Roxy the Robot) were read by an AI system. Now, with Intercom’s new AI features, all messages sent via chat or support email, whether to a human or the bot, will go through Intercom’s AI services.

As the data controller responsible for processing this data, we’ve completed a Data Protection Impact Assessment (DPIA) to assess the risks and ensure appropriate safeguards are in place. We’ve also updated our Privacy Policy to reflect this. This article explains what’s processed, why, who’s involved, and how we protect your data.


What data is processed?

AI features process:

  • Messages you send to Famly via the support chat or support@famly.co

  • Our responses

  • Help Centre articles and other Famly content used to assist with answers

If your message includes personal data (such as names of children or staff, or parent emails), that information will be included in what the AI sees. This is why we recommend only including personal data if it’s strictly necessary to resolve your issue.

We do not recommend sharing sensitive data, such as medical or safeguarding information, through support channels.


What legal basis do we rely on, and why do we collect the data?

Under data protection law, Famly acts as the data controller only with respect to the processing of support messages by AI systems. This specific processing is carried out based on our legitimate interests, which are to continuously improve our support services to meet evolving customer expectations, ensuring consistent support quality across teams through ongoing monitoring and evaluation, and developing internal insights that promote efficient and responsive service delivery.

We process messages this way to continuously improve our support services to meet evolving customer expectations, ensuring consistent support quality across teams through ongoing monitoring and evaluation, and developing internal insights that promote efficient and responsive service delivery.

We do not use this data for marketing or product profiling. AI processors are contractually prohibited from using support data for model training or any other secondary purpose.


Who processes the data?

All messages sent through support chat and email are handled by Intercom R&D Unlimited Company (based in Ireland), who process data using the following AI providers (sub-processors):

  • OpenAI

  • Anthropic

  • AWS (via Bedrock)

  • Google (Vertex AI)

These AI system providers are based in the United States. Transfers of data outside the UK and EU are covered by Standard Contractual Clauses (SCCs), with relevant addenda where needed.


How long is data kept?

The AI processors just need to use your data to deliver the output to us. They do not otherwise store the data.


What about personal data from parents or staff?

In some cases, support queries may include personal data about others. For example, the name of a parent or staff member involved in a billing or access issue.

These individuals likely won’t know their data is being included in a support request and then processed by AI tools. While this is sometimes necessary for support, we advise being cautious about what’s included in messages and limiting identifiable data unless it’s required to resolve your issue.


What safeguards are in place for your data?

  • Zero retention policy: AI processors delete inputs and outputs immediately after processing

  • No training or secondary use: Data is never used to improve or retrain AI models

  • Contractual safeguards: All transfers to the US are protected by SCCs with additional measures

  • Security controls: Intercom uses prompt protection, access controls, auditing, and regular security testing

  • Help Centre transparency: We’re publishing this article to help you and your team understand how AI is used and what it means for data shared with support


What should you do?

We recommend the following for all staff using Famly support:

  • Avoid including personal data unless it’s necessary for resolving the issue

  • Never include sensitive data (e.g. medical, safeguarding, or personal disclosures)

  • Let your team know that chat and email support messages may be processed by AI providers


Learn more

If you have any questions about how we use AI or how your data is handled, please get in touch with our support team or your Customer Success Manager.

Did this answer your question?