AI Transparency Statement

Kensley uses artificial intelligence technology to help planning professionals research, analyse, and prepare planning documents. This statement explains how we use AI, what data is processed, and the limitations you should be aware of.

1. How Kensley Uses AI

1.1Document Generation

When you use the "Generate" feature, Kensley assembles verified planning data — including site constraints, NPPF policy references, precedent applications, and your consultant notes — into a structured prompt. This prompt is sent to a large language model (LLM) which generates a draft planning document.

The LLM we currently use for document generation is Llama 3.3 70B, hosted by Groq Inc. We may change or add AI providers in the future to improve quality and performance.

1.2Research Assistant ("Ask Kensley")

When you ask a question using "Ask Kensley", your question is combined with relevant excerpts from our indexed planning corpus (officer reports, decision notices, policy documents, case law) and sent to an AI model to generate a cited answer.

1.3Site Analysis and Scoring

The reasoning engine that scores policy relevance, precedent strength, and planning confidence is a deterministic system — it uses fixed algorithms and weights, not AI. The scores you see on policy cards and precedent cards are calculated by our own scoring engine, not generated by an AI model.

However, the risk flag descriptions and conflict explanations displayed alongside scores may be generated or enhanced by AI.

2. What Data Is Sent To AI Providers

When you use an AI Feature, the following data is transmitted:

SENT:

• Postcode and site address
• Proposal description text
• Application type
• Selected policy references and their text
• Selected precedent application references and officer report extracts
• Consultant notes you have entered
• Site constraint summary

NOT SENT:

• Your name, email, or account details
• Your payment information
• Your IP address or device information
• Data from other users
• Your browsing history on Kensley

3. How AI Providers Handle Your Data

3.1Our AI providers (currently Groq Inc.) operate under Data Processing Agreements that require them to:

• Process your data solely to generate the requested output
• Not use your inputs to train, fine-tune, or improve their models
• Not retain your inputs after the response has been delivered
• Not share your inputs with any third party
• Implement appropriate security measures to protect data in transit and at rest

3.2Data is transmitted using encrypted connections (TLS 1.2 or higher).

3.3Processing typically completes within 5-30 seconds. No persistent copy of your inputs is retained by the AI provider.

4. Limitations Of AI-Generated Content

4.1AI-generated content may contain errors, inaccuracies, or omissions. Large language models can:

• "Hallucinate" — produce plausible-sounding but factually incorrect statements
• Cite policy paragraphs or case references that do not exist
• Misinterpret the relationship between planning policies
• Omit relevant considerations
• Present information out of context

4.2We mitigate these risks by:

• Providing the AI model only with verified data from our own databases
• Instructing the AI to cite sources for every claim
• Flagging gaps and unverified information with [CONSULTANT NOTE] markers
• Scoring confidence levels based on data completeness
• Using a deterministic reasoning engine (not AI) for policy and precedent scoring

4.3Despite these safeguards, AI-generated content is NOT guaranteed to be accurate, complete, or suitable for any particular purpose.

5. Your Responsibilities

5.1All AI-generated documents are draft outputs intended for professional review.

5.2You must:

• Review all generated content before use
• Verify policy citations against the original NPPF, local plan, and legislation
• Verify precedent references against the original council records
• Check that the generated document accurately reflects the specific circumstances of your site and proposal
• Add professional judgement, site-specific knowledge, and observations that the AI cannot provide
• Ensure compliance with any professional code of conduct that governs your practice

5.3You must not:

• Submit AI-generated documents to a local planning authority, the Planning Inspectorate, or any other body without professional review
• Rely on AI-generated confidence scores as a guarantee of planning outcome
• Represent AI-generated content as the definitive professional opinion of a qualified planning consultant without appropriate review and amendment

6. Professional Conduct

6.1If you are a member of the Royal Town Planning Institute (RTPI) or other professional body, you should consider whether your professional code of conduct requires disclosure of AI assistance in the preparation of planning documents.

6.2Kensley is designed to support professional judgement, not replace it. The Service does not provide legal or planning advice, and the use of AI-generated content does not create a professional advisory relationship between Kensley and any party.

7. Continuous Improvement

7.1We continuously improve our AI Features. This may include:

• Changing or adding AI model providers
• Updating system prompts and instructions
• Expanding the planning data corpus
• Refining the reasoning engine scoring weights
• Adding new document types

7.2Changes to AI Features are made to improve accuracy and usefulness. We will notify users of material changes to AI providers or data processing practices.

8. Contact

For questions about our use of AI, please contact us at ai@kensley.ai.