Skip to main content

AI Policy

How we use robots and ai to help service what we do for you

Version 1.0 | Effective December 2025

Our approach to AI

At Shapes we believe in using the latest technology and embracing the future of applications and the web. Artificial Intelligence isn’t just a tool for us but part of how we deliver smarter, faster and more creative solutions for our clients.
We’re equally committed to transparency, data protection and maintaining the human expertise that makes our work truly valuable.

This policy explains how we use AI across our business, what safeguards we have in place and what it means for our clients.

1. Why we use AI

We use AI to enhance – not replace – our team’s expertise. It helps us to :

  • Work more efficiently on research, analysis, and initial production outputs and creative concepts
  • Generate insights faster from large datasets and complex information
  • Prototype and test ideas more quickly during the design and development process
  • Maintain consistency in brand voice and creative output
  • Stay competitive by leveraging the best tools available

What AI never replaces: Your project and our work for you at Shapes always has human oversight, creative direction and quality control.
All advice whether creative, strategic or financial is only ever given by our team and not by AI tools.
All AI input to anything we do is verified and overseen by our team before reaching your eyes. 

2. AI tools we use

We’re transparent about the technology in our toolkit:

Creative & Design Tools

  • Claude AI & ChatGPT – For initial ideas, copywriting support, research analysis and design efficiency
  • Adobe Firefly – AI-assisted graphics and mockups that we refine for purpose
  • Canva AI – AI-assisted graphics and image manipulation that we refine for purpose
  • Nano Banana – Rapid visual prototyping and graphic creation
  • Figma AI – Generating workable prototypes and coded examples of design work, plus some clever AI image editing

Analysis & Insights Tools

  • Microsoft Clarity – UX insights and user behaviour recommendations
  • Google AI tools – Embedded analytics and pattern recognition
  • Claude AI – Analysing datasets, identifying themes and generating content direction

Business Management Tools

  • Xero – Accounting software with AI-powered task efficiency
  • Vinyl – Meeting recording with AI-generated summaries, action points and client communication recommendations
  • Karbon HQ – Practice management with AI capabilities for workflow efficiency
  • Google Business Suite – We run most of our admin side of our business through Google. Google has a variety of AI features that enhance efficiency.
  • Briefcase – Bookkeeping and financial close for speed, consistency and VAT compliance insight

Project Management

  • Claude AI – For project documentation, common pattern identification, and workflow optimisation

3. How we use AI in your work

Here’s what AI involvement looks like in practice:

Creative Work (Brand, Design, Web, Marketing)

  • Brainstorming & ideation – AI helps our team generate initial concepts and turning sketches into digital forms
  • Research – Quickly analysing competitor landscapes, audience insights and industry trends to give us simple overviews and support of our research
  • Copywriting – Creating first drafts and maintaining tone of voice consistency
  • Design mockups – Generating quick visual concepts that our designers then refine
  • Code generation – Producing functional prototypes and website components
  • Content analysis – Identifying themes from user research and stakeholder interviews

What you receive: Every deliverable is refined by humans, quality-checked and aligned with your brief. AI outputs are always our starting point, never our end product (unless that was the brief!).

Accounting & Financial Services

  • Data processing – Speeding up routine bookkeeping tasks (via Xero and other tool’s built-in AI)
  • Pattern recognition – Flagging unusual transactions or identifying trends

What AI doesn’t do: Provide financial advice. All recommendations, tax guidance and strategic financial advice and support comes from our qualified team members.

Client Communication & Project Management

  • Meeting notes – Vinyl AI captures key points, action items and decisions
  • Documentation – Organising project assets and maintaining clear records
  • Workflow optimisation – Karbon HQ’s AI helps us stay on top of deadlines and deliverables

4. Data Protection & Privacy

We take your data seriously. Here’s how we protect it:

Financial & Accounting Data

Maximum security approach:

  • Client financial data is processed exclusively through Xero (which is UK GDPR compliant, ISO 27001 certified and SOC2 compliant)
  • We use Claude for Work (business accounts) for any AI-assisted analysis, which guarantees:
    • Zero data training under any circumstances
    • Minimal data retention by default
    • Commercial-grade Data Processing Agreements (DPAs)
    • Full UK GDPR compliance
  • Our Claude for Work account operates under Commercial Terms of Service, providing the highest level of data protection available

What this means: Your financial information never contributes to training AI models and is protected by enterprise-grade security measures.

Creative & Strategic Data

When using AI tools for brand briefs, strategy documents or user research:

  • We use Claude for Work business accounts with guaranteed zero external data training
  • We anonymise sensitive information before input where appropriate
  • All tools are configured for maximum privacy settings
  • We never input information that could identify individuals or reveal commercially sensitive details
  • For other AI tools (ChatGPT, Adobe Firefly), we use paid versions with training opted out where applicable

Meeting Recordings (Vinyl)

  • Recordings are used solely for our team’s internal reference and client communication
  • AI-generated summaries are reviewed before sharing
  • Recordings are stored securely and deleted according to our retention policy
  • Clients can request not to be recorded at any time

Data Retention

  • Financial data: Retained per UK tax and accounting regulations (minimum 6 years)
  • Claude for Work interactions: Minimal retention under Commercial Terms (significantly shorter than consumer accounts)
  • Project files: Retained per our standard client agreements.
  • Meeting recordings: Retained per our standard client agreements.

5. UK GDPR Compliance

All our AI usage complies with UK GDPR requirements:

  • Lawful basis: We process data based on legitimate interests and contractual necessity
  • Data minimisation: We only input what’s necessary for the task
  • Purpose limitation: Data is only used for your specific project
  • Storage limitation: We don’t keep data longer than needed
  • Right to erasure: Clients can request deletion of their data at any time

Our AI tools are configured to align with these principles and we maintain written policies for team training and compliance monitoring.

6. What we don’t do with AI

To be absolutely clear, here’s what AI will never do at Shapes:

❌ Replace human judgment on creative direction or strategic decisions
 ❌ Provide accounting advice under any circumstances
 ❌ Make final decisions about your brand, design, or financial matters
 ❌ Train external AI models on your confidential data (we’ve opted out)
 ❌ Share your information with third parties or AI training datasets
 ❌ Generate final deliverables without human refinement and quality control

7. Quality & Accuracy

Every AI-generated output goes through our quality process (where relevant for the task at hand) : 

  1. Human review – Team members check for accuracy, tone, and relevance
  2. Brand alignment – Ensuring outputs match your creative vision and requirements
  3. Fact-checking – Verifying any data, statistics, or claims
  4. Legal compliance – Confirming all work meets regulatory standards (especially for accounting)
  5. Client approval – You always have final say on deliverables

8. Disclosure & Transparency

What we tell you

We’ll proactively disclose AI involvement when:

  • It significantly shaped the creative direction or output
  • You specifically ask about our process
  • It’s relevant to understanding the deliverable

What we’re telling you now

This policy itself serves as disclosure that AI tools are integrated into our workflow across creative, analytical, and administrative functions. If you ever want to know more about how AI was used in your specific project, just ask any of the team, we’re more than happy to explain.

9. Client Choice & Control

You’re always in control:

  • Opt-out option: If you prefer us not to use AI tools on your project, let us know upfront
  • Meeting recording consent: We’ll always ask before recording, and you can decline
  • Data requests: You can request details about what data we’ve processed and ask for deletion
  • Questions welcome: We encourage dialogue about our AI usage at any time

10. Intellectual Property

When AI assists in creating work for you:

  • Ownership: You own all final deliverables as outlined in our standard contracts
  • AI-generated components: We ensure all outputs are refined enough to be considered original work
  • Copyright: We check that AI-generated content doesn’t infringe on existing copyrights
  • Licensing: All AI tools we use permit commercial use of outputs (with human modification)

11. Continuous Improvement

AI technology evolves rapidly. We commit to:

  • Regular policy reviews – Updating this document as tools and regulations change
  • Team training – Ensuring everyone understands best practices for AI and data protection
  • Tool evaluation – Continuously assessing whether our AI tools meet privacy and quality standards
  • Client communication – Notifying you of any significant policy changes

12. AI Ethics & Responsibility

We’re thoughtful about AI’s broader impact:

  • Human-centred approach – AI enhances human creativity, it doesn’t replace it
  • Bias awareness – We review AI outputs for potential bias and stereotyping
  • Environmental consideration – We balance AI’s efficiency benefits against its energy consumption. We also offset some of our energy use through more purchases of trees through Ecologi. 
  • Fair and transparent attribution – We don’t claim AI-generated ideas as purely human-created
  • Job security – AI allows our team to focus on higher-value work, not to reduce headcount

ICAEW Professional Ethics & AI Governance 

As a provider of accounting services, we align our AI practices with the ICAEW Code of Ethics and the IESBA (International Ethics Standards Board for Accountants) fundamental principles. These five core principles guide how we use AI responsibly and ethically in our work:

Integrity – We are transparent and honest about our use of generative AI, especially when it supports research, analysis, or advice delivered to clients. We validate the integrity of AI model inputs before use, and we implement independent reviews to ensure AI isn’t being prompted to justify pre-determined conclusions. Our team never uses prompt engineering to deliberately produce biased outputs.

Objectivity – We recognise that AI tools can introduce bias and we’re aware of their data sources and limitations. We exercise professional judgement rather than over-relying on AI outputs, avoiding automation bias that could compromise our recommendations. For financial advice, client data, and strategic decisions, our qualified team members always apply independent thinking and professional scepticism to AI-generated insights.

Professional Competence and Due Care – AI tools don’t replace qualified accountants. We ensure our team understands how each AI tool works, including its limitations and risks. We remain aware that public AI models may lack current legislative developments and may contain inaccuracies. (See section 11 for our commitment to continuous professional development in AI use.)

Confidentiality – We treat client and organisational information with strict confidentiality. (See section 4 for full details on how we protect your data and the safeguards we currently have in place.)

Professional Behaviour – We avoid overstating or misrepresenting AI outputs and don’t claim AI-generated concepts as purely human-created. We stay informed about relevant legal frameworks around AI (data protection, intellectual property, employment law) and emerging AI-specific regulations. Should we identify any use of AI that breaches or risks breaching laws and regulations, we fulfil our duty to act in the public interest in line with IESBA’s NOCLAR (Non-Compliance with Laws and Regulations) obligations.

We also ensure that we never assume management responsibility for client AI systems or decision-making. Our role is to advise on and implement appropriate controls. Responsibility for AI governance and decisions always remains with the client. This protects both parties and ensures we maintain professional independence.

13. What this means for your project

In practice, our AI policy ensures:

✅ Your data is protected to the highest UK GDPR standards
 ✅ Financial information never trains AI models
 ✅ All advice and deliverables are human-verified
 ✅ You understand how AI contributes to your project
 ✅ You can opt out if you’re not comfortable
 ✅ We’re transparent about our processes
 ✅ Quality and accuracy are never compromised

14. Questions or Concerns?

We’re always happy to discuss our AI approach in more detail. If you have questions about:

  • How AI is being used on your specific project
  • Data protection measures for your information
  • Opting out of AI tools
  • Technical details about our tools and safeguards

Just get in touch with your Shapes team contact, and we’ll provide clear answers.

Contact & More Information

Shapes
 Website: shapes.team
 Email: [email protected]

Document Control:
 Policy Owner: Jim Leeves / Steve Kreeger
Last Reviewed: December 2025
 Next Review: June 2026

This policy demonstrates our commitment to using AI responsibly, transparently, and always in your best interests. We’re excited about the possibilities AI brings to our work but never at the expense of quality, privacy, or the human touch that makes Shapes distinctively Shapes.