Human Oversight
AI should assist people, not replace them. Important decisions should involve appropriate human judgement.
AI Governance
Responsible AI Policy
Artificial intelligence has the potential to improve productivity, accelerate decision-making and support better governance. At Aqitas, we believe these benefits must be balanced with transparency, accountability and human oversight.
Our approach is guided by the principle that AI should support informed human decision-making rather than replace it.
01 / Purpose
Building trust through responsible artificial intelligence.
This Responsible AI Policy explains how Aqitas designs, develops and deploys AI-enabled capabilities, and the principles that guide our use of artificial intelligence across our platform and professional services.
The policy establishes a responsible framework for AI-enabled services provided by Aqitas. Its objectives are to promote trustworthy AI, support transparent decision-making, protect customer information, reduce operational and governance risk, encourage meaningful human oversight, improve consistency and accountability, and align with evolving legal and regulatory expectations.
02 / Scope
Where the policy applies
This policy applies to the Aqitas SaaS platform, AI-assisted due diligence workflows, governance automation, evidence analysis, document summarisation, risk assessment support, professional and consultancy services where AI is used, and internal AI use by Aqitas personnel where customer information is processed.
Third-party AI services integrated into the Aqitas platform are also expected to meet appropriate security and privacy standards.
03 / Principles
The standards that guide responsible AI at Aqitas
Transparency
Users should understand when AI has contributed to an output or recommendation.
Accountability
AI-generated recommendations do not transfer accountability from decision-makers to software.
Fairness
We seek to minimise unnecessary bias and support fair, evidence-based decision-making.
Privacy
Customer information is processed only for the purposes requested by the customer.
Security
Responsible AI requires secure systems that protect customer information.
04 / Human decision-making
AI supports governance workflows rather than replacing judgement.
AI within Aqitas is designed to support governance workflows rather than automate consequential decisions without oversight.
- Summarising documents.
- Identifying potential risks.
- Highlighting missing evidence.
- Suggesting workflow actions.
- Drafting governance records.
- Organising information.
Customers should independently assess AI-generated outputs before relying upon them. Where decisions have significant legal, financial or operational consequences, appropriate human review should always be applied.
05 / Governance
Responsible AI is an operating model, not a one-time control.
As Aqitas grows, we intend to continue developing policy review, internal oversight, risk assessment, AI feature evaluation, security review, customer feedback, documentation and operational monitoring.
06 / Customer data
Customer data remains customer-owned.
Unless otherwise agreed, customer information is processed solely to deliver the requested functionality, uploaded documents are not retained after processing where retention is not required for the agreed service, customer information is not sold or shared for advertising purposes, and customer information is not used to train Aqitas AI models or third-party foundation models.
These commitments apply regardless of whether AI-assisted functionality is enabled.
07 / AI technologies
Carefully selected capabilities for specific governance tasks.
Aqitas may utilise carefully selected AI technologies to support natural language processing, document summarisation, information extraction, evidence organisation, workflow automation, risk identification and governance recommendations.
Technology providers are assessed as part of Aqitas' broader security, privacy and governance processes.
08 / Explainability
AI outputs should connect back to accountable context.
Documents, records and submitted material linked to the output.
Workflow actions and review stages provide context.
Assisted summary, signal or recommendation.
Named accountability for review and approval.
Decision record and approval timestamp retained for governance.
09 / Accuracy, fairness and bias
AI outputs should be reviewed, not treated as definitive.
Artificial intelligence is probabilistic. AI-generated outputs may contain inaccuracies, omit relevant information, produce incomplete responses, misinterpret context or reflect limitations within training data.
Aqitas recognises that AI systems may reflect or amplify bias. We seek to reduce this risk by encouraging evidence-based workflows, maintaining human oversight, reviewing system behaviour during development, monitoring customer feedback and continuously improving platform capabilities.
10 / Security
Responsible AI requires secure systems.
AI functionality is supported by security measures including encryption in transit, encryption at rest, identity and access management, authentication controls, audit logging, infrastructure monitoring and secure software development practices.
Security controls continue to evolve as our platform develops.
This policy will evolve with technology, regulation and customer needs.
Aqitas will periodically review this policy to reflect technological developments, customer feedback, emerging risks, regulatory expectations, industry best practice and platform enhancements.
- Document
- Responsible AI Policy
- Version
- 1.0
- Status
- Active
- Effective Date
- July 2026
- Applies To
- Aqitas website, platform and consultancy services
Questions about this policy?
Questions regarding this Responsible AI Policy may be directed to Aqitas Limited.
[email protected]