AI Usage in Umbraco Support

Our Human-Centered Approach to AI

At Umbraco, we believe artificial intelligence can enhance the customer experience when used responsibly and thoughtfully. Our approach to AI is guided by a simple principle: technology should support people, not replace them.

We use AI-powered tools to help customers access information more quickly, reduce friction in support interactions, and provide assistance outside standard business hours. These tools are designed to complement the expertise of our support teams and the knowledge shared by the Umbraco community.

As a company, we value trust, transparency, collaboration, and human expertise. While AI can help solve many questions efficiently, we recognize that some situations require deeper technical knowledge, human judgment, empathy, or a more personalized approach. For this reason, our support specialists remain an essential part of the customer experience and are available to assist when needed.

We continuously monitor the quality, accuracy, effectiveness, and customer satisfaction of our AI-assisted support services. Customer feedback plays a critical role in helping us improve these systems and ensuring they continue to deliver value while meeting the standards our customers expect.

Our goal is not to automate support. Instead, we aim to create a better overall experience by combining the strengths of AI technology with the expertise, creativity, and care of our people.

The following disclosure explains how AI may be used within our Customer Experience services, what information may be processed, and the safeguards we have in place to support responsible and transparent use of AI.

AI Systems Used in Support

Umbraco uses approved AI systems in limited and controlled scenarios.

Umboto (AI Support Assistant)

Umboto is Umbraco's AI-powered support assistant. It is designed to help customers find information, troubleshoot issues, and receive guidance related to Umbraco products and services.

Umboto is delivered through Intercom (Fin) and combines multiple technologies and workflows to support customer interactions.

Some responses are provided through predefined support workflows and content created and maintained by Umbraco teams. These workflows are designed to deliver consistent guidance for common questions and support scenarios.

Umboto also uses Fin AI, Intercom's AI support platform, which generates responses based on available knowledge and context. To ensure answers are relevant and accurate for the Umbraco ecosystem, the AI experience is configured and continuously improved by Umbraco's support team. This includes custom instructions, support processes, documentation, and Umbraco-specific content.

As a result, Umboto's responses may be generated through a combination of predefined workflows, AI-generated content, and human-authored knowledge resources. The support team continuously monitors performance and customer feedback to improve the quality, accuracy, and usefulness of the assistance provided.

While Umboto can independently resolve many questions, it is intended to complement - not replace - human support. When additional expertise, context, or judgment is required, customers may be directed to, or assisted by, members of the Umbraco team.

As with any AI system, responses may occasionally be incomplete or inaccurate, and customers should exercise appropriate judgment when acting on AI-generated information.

How Umboto Uses Your Information

Support conversations submitted to Umboto may be used to:

• Generate responses within the current support interaction

• Improve the quality and accuracy of Umboto

• Support human review processes for quality assurance and chatbot improvement

Human Review

Some conversations handled by Umboto may be reviewed by Umbraco employees to:

• Improve chatbot performance

• Validate response quality

• Investigate support issues

• Optimize support workflows

Not all conversations are reviewed manually.

Claude (Anthropic)

In specific troubleshooting scenarios, Umbraco support engineers may use Anthropics Claude to help analyze:

• Error logs

• Stack traces

• Configuration issues

• Small, relevant code excerpts

• Technical debugging information

This is done only when necessary to assist with resolving a support issue more efficiently.

Important Safeguards

When Claude is used:

• Only relevant information is shared

• The scope is limited to the investigation

• The data is not used to train Anthropic’s models when using approved enterprise or paid services

• The information is not retained for unrelated future use or product development

Umbraco employees are instructed to avoid sharing unnecessary personal data or unrelated confidential information.

Gemini (Google)

Gemini may be used internally to summarize support conversations when a case is escalated from Umboto to a human support engineer.

This helps reduce resolution time and improves continuity between AI-assisted and human-assisted support.

Your Code, Data, and Confidential Information

Your Code Is Not Used to Train Public AI Models

If you share source code, configuration files, or technical information during a support case:

• It is used only for the purpose of investigating and resolving the support issue

• It is not intentionally used to train public AI models

• It is not reused to develop unrelated Umbraco products or features

• Access is limited to approved tools and authorized personnel

Where AI-assisted analysis is used, Umbraco aims to minimize the amount of customer information processed.

Please Avoid Sharing Sensitive Information

We recommend that customers do not include unnecessary sensitive information in support conversations, including:

• Passwords

• API keys

• Secrets or tokens

• Credit card information

• National identification numbers

• Personal health information

• Unrelated personal data

If sensitive information must be shared for troubleshooting purposes, please limit it to what is strictly necessary.

Right to be forgotten

Individuals whose personal data has been processed by Umbraco's AI systems have the right to request erasure of that data in accordance with Article 17 of the GDPR. Upon receiving a valid erasure request, Umbraco will. Requests can be submitted here following the written procedure.

Project Access and Customer Environments

In some support cases, customers may provide Umbraco with access to development, staging, or production environments to assist with troubleshooting.

In these situations:

• Umbraco support personnel remain responsible for all investigation activities

• AI tools may assist with analysis of technical findings, logs, or diagnostics

• AI tools are not used to independently access customer systems

• AI tools are not authorized to make code changes or deploy changes to customer environments

Customers remain responsible for ensuring they have appropriate permissions and legal authority to share environment access or customer data with Umbraco.

Data Protection and GDPR

Umbraco evaluates AI usage in accordance with applicable data protection and security obligations, including GDPR requirements where applicable.

Our approach includes:

• Use of approved AI vendors and services

• Restricted internal usage policies

• Human oversight

• Reviewing vendor commitments regarding data retention and model training

Where AI providers act as subprocessors, Umbraco evaluates contractual and security safeguards before approval for internal use.

Approved AI Tools and Governance

Umbraco employees may only use AI tools that have been approved internally for support-related activities.

Internal governance measures may include:

• Access controls

• Internal usage policies

• Security and legal review

• Employee guidance and training

Unauthorized use of non-approved AI systems for customer support activities is prohibited.

Umbraco Develops TIA/ DPIA

We maintain these TIA/DPIA documents internally. While they are not deemed suitable for public release, they can be made available upon request to gdpr@umbraco.com