How we use AI — and what we never let it replace
We use internal intelligence tools to help our people organise information, prepare work and identify matters requiring attention. Technology supports those activities. Material customer-impacting decisions remain subject to authorised human oversight and the applicable client mandate.
Our approach to artificial intelligence distinguishes between tools that support human work and the human authority required for material decisions. AI supports analysis and operational efficiency; it does not replace accountable human judgement.
AI can assist judgement. It does not inherit authority.
As artificial intelligence tools become more capable, organisations face genuine governance risks — including over-reliance on automated recommendations, biased or disproportionate outputs, inaccurate generated information, inappropriate use of personal data, unclear responsibility for consequential actions, or automation exceeding the client mandate.
Internal intelligence tools are intended to operate inside defined purposes, permissions and workflow boundaries. They may support analysis, organisation, preparation, monitoring or recommendations. They do not replace the accountable person responsible for exercising the applicable authority.
Assistance across approved operational workflows.
Depending on the approved workflow, internal tools may support our personnel in specific operational activities:
Tools may organise account information according to authorised portfolio and workflow parameters so personnel can review relevant information more efficiently. Any prioritisation or recommendation remains subject to applicable client parameters, governance controls and human operational authority.
Tools may assist with drafting or organising communication material within approved frameworks. Communications remain subject to applicable authorised workflows and human oversight requirements.
Tools may identify patterns, exceptions or conditions that warrant human attention. A flag is an input for review, not a finding of liability or misconduct.
Tools may assist in organising operational or compliance information for internal or client review.
Assistance has a boundary.
Finder & Partners sets strict operational boundaries around technology. Internal AI tools are not authorised to independently:
Finder & Partners does not receive, collect, hold or process customer payments. AI systems therefore do not accept or process customer payments. Payments must be made directly through the authorised channels of the appointing client.
Operational customer information must not be used to train general AI models or for unrelated model development unless a separately authorised, legally permissible and governed process has been approved.
Naya provides guidance. It does not become the account system.
Naya is UnitKredit's public assistance and guidance interface. It is designed to assist members of the public, account holders and corporate prospects by providing approved information and routing enquiries.
Naya does not retrieve account-specific operational information through its public knowledge pathway.
Account-specific information (balances, status, history, dispute records) remains isolated within separate operational environments.
Naya is designed to ground responses in approved information and defined access boundaries. Where approved information is unavailable or a request falls outside the permitted scope, the intended behaviour is to decline, clarify or route the enquiry rather than retrieve restricted information. Generated responses may still require human review where appropriate.
Technology does not become the accountable party.
Responsibility for operational decisions remains with authorised personnel acting within their roles and the applicable mandate.
AI-generated recommendations, summaries or drafts do not transfer accountability from the person exercising the relevant authority to the system that assisted them.
Where decisions or actions require an approval record, that record should identify the responsible human role or person according to the applicable operational control.
Human oversight matters, but it is not the only fairness control.
Use AI only for a documented operational purpose.
Use information relevant to that purpose.
Personnel should be able to question and override tool outputs.
Where appropriate, review patterns for anomalous, disproportionate or unintended outcomes.
Material changes to models, prompts, parameters or automation should be evaluated before broader deployment.
Questionable outputs should be capable of being escalated to authorised personnel.
Protected or sensitive characteristics should not be used to drive operational treatment unless their use is lawful, necessary for an authorised purpose and specifically governed.
People should be able to ask how technology is involved.
Account holders and clients may raise questions about how technology supports Finder & Partners' recovery activities. Naya can route responsible-AI or transparency enquiries into the appropriate controlled process.
- • general role of technology in the relevant process
- • whether tool assistance formed part of the workflow
- • the role of human oversight
- • available review or enquiry pathways
- • internal prompts and system instructions
- • security architecture details
- • restricted model parameters
- • third-party confidential information
- • unverified account data
Our operational commitments.
MALAYSIA AI GOVERNANCE CONTEXT
Finder & Partners considers Malaysia's National Guidelines on AI Governance and Ethics (AIGE) as one relevant reference point for responsible AI governance.
Ask how technology supports our work — and where its authority ends.
Ask Naya to identify the appropriate pathway for:
Account-specific enquiries require the appropriate controlled identity and account-authorisation process.
This statement describes the responsible-AI principles applied to AI-assisted tools and public assistance channels used by Finder & Partners Sdn Bhd under the UnitKredit digital identity.
It does not expand Naya's authority or create access to operational customer information.
Our use of AI and related governance controls may evolve. Material changes to the role or authority of AI in our operations should be reflected in this statement.