Ai in Canvas
Enterprise AI designed to enhance customer insight, not replace human judgement
Canvas uses Artificial Intelligence (AI) to help organisations understand customer feedback faster and more effectively. Our AI capabilities assist with summarising information, identifying trends, analysing sentiment and translating feedback, enabling users to spend less time producing reports and more time acting on insights.
AI within Canvas is designed as an assistive technology. It supports decision-making but never replaces human judgement. Responsibility for interpreting AI-generated outputs and making business decisions always remains with the user.
Our AI Principles
The use of AI within Canvas is guided by five core principles:
AI supports human decision-making and does not make autonomous business decisions.
Customer data is protected using enterprise-grade security and privacy controls.
AI usage is transparent, with AI-generated content clearly identified where appropriate.
Human review remains essential before AI-generated outputs are relied upon.
AI capabilities operate within Watermelon Research's ISO/IEC 27001:2022 certified Information Security Management System (ISMS).
How AI is used within Canvas
Canvas currently uses AI to assist with a range of customer experience and Voice of Customer activities, including:
Dashboard Summaries
Generate concise summaries of customer feedback, trends, satisfaction metrics and emerging issues.
Report Drafting
Assist users by producing draft executive summaries, commentary and insight narratives that can be reviewed and refined before publication.
Text Analytics
Analyse large volumes of customer comments through:
Language translation
Theme and topic identification
Sentiment analysis
Trend detection
Insight recommendations
These capabilities are designed to accelerate analysis while maintaining human oversight throughout the reporting process.
What AI is not used for
Canvas AI is not designed or intended for:
Automated customer decision-making
Credit or lending decisions
Insurance or eligibility assessments
Fraud determinations
Legal or regulatory decisions
Compliance determinations
Any autonomous decision that could materially affect an individual
AI outputs should never be relied upon as the sole basis for significant business or customer decisions.
Human Oversight
All AI-generated outputs are advisory in nature.
Users remain responsible for:
validating AI-generated summaries
reviewing supporting data
applying professional judgement
ensuring outputs are appropriate before distribution or decision-making
Where reports are shared externally or used to support strategic or operational decisions, human review should always take place.
Technology Platform
Canvas AI capabilities are powered through Microsoft Azure OpenAI Service.
This provides enterprise-grade security, authentication and privacy controls while enabling advanced natural language processing capabilities.
Customer prompts and data submitted through Azure OpenAI are not used to train OpenAI foundation models.
Privacy and Data Protection
Watermelon Research is committed to protecting customer information throughout all AI-assisted processing.
Our AI capabilities are designed to support compliance with applicable privacy legislation, including:
UK GDPR
EU GDPR (where applicable)
Australian Privacy Act 1988 and Australian Privacy Principles (APPs)
Where practical, organisations should minimise the inclusion of personally identifiable information within AI prompts and reporting workflows.
Security and Governance
AI capabilities operate within Watermelon Research's ISO/IEC 27001:2022 certified Information Security Management System.
Security controls supporting AI services include:
role-based access controls
least-privilege access
encryption in transit
encryption at rest
audit logging
change management
supplier assurance
risk management
security monitoring
incident management
AI functionality is subject to the same governance, security and operational controls as the wider Canvas platform.
Responsible AI
Like all generative AI technologies, AI-generated content may occasionally contain inaccuracies or incomplete information.
To reduce this risk, Canvas applies a combination of technical controls and operational governance, including:
Incorrect summaries
AI-generated summaries should always be reviewed and validated against the underlying customer feedback before being relied upon.
Hallucinations
Users are responsible for verifying AI-generated content and ensuring it accurately reflects the source information.
Bias or misinterpretation
Human oversight and contextual review help ensure outputs are interpreted appropriately and fairly.
Privacy exposure
Data minimisation, secure processing and appropriate handling of personal information help reduce privacy risks.
Security threats
Enterprise-grade Azure security controls, access management and continuous monitoring help protect AI processing.
Over-reliance
AI is positioned as an assistive capability. Responsibility for interpreting outputs and making decisions always remains with authorised users.
Data Residency
Canvas is deployed within Microsoft Azure.
AI processing takes place within the Azure region configured for the relevant Canvas deployment, subject to Microsoft Azure OpenAI regional availability.
Regional deployments are available for customers operating within the:
United Kingdom
European Union
Australia
Frequently Asked Questions
Is our data used to train AI models?
No. Customer data submitted through Canvas AI capabilities is not used to train OpenAI foundation models.
Does AI make decisions on our behalf?
No. AI assists with summarisation, reporting and analysis only. All decisions remain the responsibility of authorised users.
Can AI-generated content be trusted?
AI provides a valuable starting point for analysis but should always be reviewed and validated before use.
Is AI secure?
Yes. AI capabilities are delivered using Microsoft's Azure OpenAI Service and operate within Watermelon Research's ISO/IEC 27001:2022 certified Information Security Management System.
Can AI features be disabled?
Where required, AI capabilities can be configured or disabled to meet individual client requirements.
Our Approach
Watermelon Research believes AI should make customer insight faster, more accessible and easier to understand—without compromising privacy, security or governance.
Canvas uses AI as an assistive capability that enhances human expertise rather than replacing it. By combining enterprise-grade security, robust governance and responsible AI practices, organisations can confidently use AI to unlock greater value from their customer feedback while maintaining full control over their data and decision-making.