AI can support better thinking. It should not replace human judgement.
Pridmora uses AI to support manager development, reflection, preparation and the recognition of potential development patterns. Our approach is built around human oversight, appropriate evidence, privacy boundaries and clear limits on what AI should be used to decide.
AI-supported outputs can help users think, notice, organise and interpret. They should not be treated as unquestionable facts or autonomous decisions about a person.
Responsible AI begins with clear design principles.
Human judgement
People remain responsible for interpreting AI-supported outputs and deciding what action, if any, should follow.
Evidence before certainty
Development interpretations should earn confidence through relevant evidence rather than isolated activity.
Privacy by design
AI functionality operates within Pridmora's wider access, privacy and organisational visibility boundaries.
Development purpose
AI is used to support development and learning rather than covert employee surveillance or automated performance scoring.
What Pridmora's AI is designed to support.
Aurelia is Pridmora's AI-supported development capability. Depending on the feature being used and the information authorised for that workflow, Aurelia may help users:
Prepare
Structure thinking before development conversations, workplace situations or other relevant management activity.
Reflect
Explore what happened, what was learned and what might warrant further attention or development.
Organise evidence
Help identify observations, themes or possible connections within authorised development information.
Recognise potential patterns
Surface possible recurring development patterns for human review rather than silently converting AI inference into established fact.
AI-supported development should include meaningful human review.
Pridmora's development model uses human review at important points rather than assuming an AI-generated interpretation is automatically correct.
Aurelia may generate a suggestion, interpretation, draft, question or potential development pattern.
The relevant authorised user considers the output in context rather than assuming it is correct because it was generated by AI.
Where an interpretation contributes to Development Intelligence, relevant evidence, limitations and confidence should be considered.
Users remain responsible for the decisions and actions they take following AI-supported development activity.
Pridmora AI is not designed to become an autonomous decision-maker about people.
Pridmora is development infrastructure. It is not designed as an employee-surveillance, disciplinary, clinical, safeguarding or HR case-management system.
Pridmora AI should not be relied upon to:
- make autonomous recruitment or employment decisions;
- determine whether someone should be promoted, disciplined or dismissed;
- provide legal or employment-law advice;
- make safeguarding determinations;
- provide clinical or medical advice;
- replace formal HR or employee-relations processes;
- diagnose an individual's psychological or medical condition;
- covertly monitor individual employee performance; or
- turn a development signal into an unsupported judgement about capability.
Where an issue requires formal HR, People, safeguarding, clinical, legal or other specialist handling, the appropriate organisational process and professional advice should be used.
AI interpretation is not the same as established evidence.
Development is complex. A single comment, conversation or event may be relevant without being sufficient to support a confident conclusion.
Build confidence carefully
Potential themes can remain emerging while further relevant evidence develops. Repetition, context, human review and contradictory evidence can all matter when interpreting a pattern.
False certainty
AI-generated language should not make an early signal sound more certain than the underlying evidence justifies. The existence of an AI interpretation does not itself establish that the interpretation is true.
Organisation-level insight requires additional safeguards.
Pridmora's organisational Development Intelligence is intended to help authorised leaders understand collective development patterns without giving them a general right to inspect managers' private development records.
Safeguards may include:
- minimum contributor thresholds;
- suppression of below-threshold themes;
- separation of organisational oversight and practitioner permissions;
- restrictions on the underlying private information available to Organisation Leads;
- human interpretation of organisational patterns; and
- clear language where evidence is insufficient to support a stronger conclusion.
Development Intelligence should inform development investment, organisational learning and support. It should not be treated as an individual performance score or a substitute for formal people-management processes.
AI operates within Pridmora's privacy model.
Authorised information
Aurelia receives information assembled for the relevant authorised workflow. It is not given unrestricted access to everything held about a person or organisation within Pridmora.
Identity Vault separation
Identifying profile information held in Pridmora's Identity Vault is kept separate from the person-level external AI processing path. The Identity Vault is not provided to the external AI model as a general source of identity information.
Data minimisation and pseudonymisation
Before authorised person-level content is sent for external AI processing, Pridmora applies controls intended to detect and minimise identifiable information, including names and certain other identifiers, wherever feasible.
Free text can contain identifying information that automated detection does not recognise. Users should therefore avoid entering unnecessary patient-identifiable, clinical, medical, safeguarding, disciplinary, grievance or other highly sensitive case information into development workflows. Original authorised evidence may remain within Pridmora for authorised traceability and human review.
Specialist AI infrastructure may be used to provide authorised features.
Third-party AI processing
Pridmora uses specialist AI model/API infrastructure to provide defined AI-supported functionality. Customer Personal Data processed through that infrastructure is subject to Pridmora's applicable contractual, data-protection and AI governance arrangements.
Pridmora controls the information assembled for each authorised AI workflow before it is sent to the external AI service. The external AI model is not given unrestricted access to Pridmora's database or the Identity Vault.
For person-level AI processing, Pridmora applies data-minimisation and pseudonymisation controls intended to reduce unnecessary disclosure of identifiable information before external processing. These controls are designed to reduce privacy risk but are not represented as perfect anonymisation.
The use of an external AI provider does not remove Pridmora's responsibility to establish appropriate service, privacy and governance controls around the way the technology is integrated into Pridmora.
Responsible AI is a shared responsibility.
Users remain responsible for decisions and actions they take following AI-supported activity.
Users should:
- consider AI outputs critically rather than accepting them automatically;
- use professional and organisational judgement;
- consider whether relevant evidence supports an interpretation;
- avoid entering unnecessary highly sensitive information;
- correct or reject AI-supported interpretations where appropriate; and
- move formal HR, legal, safeguarding or clinical matters into the appropriate authorised process.
AI can be useful without being infallible.
AI systems can produce incomplete, inaccurate, overly confident or contextually inappropriate outputs. They may fail to understand important organisational, interpersonal or situational context.
For that reason, Pridmora does not position AI-generated content as inherently authoritative. Users should consider the source evidence, context, uncertainty and their own professional judgement.
Our AI approach will evolve with the technology and evidence.
Pridmora maintains wider governance covering AI, privacy, information security, access control, secure development, supplier management, incident response and Development Intelligence.
AI-supported functionality and this public statement may be reviewed when material functionality changes, AI providers change, new risks are identified, legal or regulatory requirements develop or evidence from real-world use indicates that safeguards should be strengthened.
Responsible AI sits within a wider governance framework.
You can find further information in our Privacy Notice, Cookie & Browser Storage Notice and Trust & Governance page.
Questions about Pridmora's use of AI?
Customers, users and organisational reviewers can contact Pridmora if they require further information about Aurelia, AI governance, privacy or Development Intelligence.
Contact Pridmora