Responsible AI and AI Governance Notice
Last updated: 17 November 2026
How we approach the responsible design, use and oversight of artificial intelligence across our technology services and business operations.
1. About this notice
Nexa Point Group Ltd (“Nexa Point”, “we”, “our” or “us”) provides managed IT, cybersecurity, cloud, infrastructure, consulting, procurement, AI and automation, AI governance, and telecoms and connectivity services. We recognise that artificial intelligence (AI) can improve how organisations work, but it can also create risks involving privacy, confidentiality, accuracy, security, fairness and accountability.
This notice describes the principles we apply when evaluating, recommending, developing or using AI-enabled tools and services, including third-party technology. It covers our own operations and relevant customer engagements. The controls and technical features available in a particular service depend on its agreed scope, supplier configuration and deployment status.
This notice supplements, but does not replace, our Privacy Policy, Cookie Policy, Recruitment Privacy Notice, customer contracts, data processing agreements and service-specific information. It is a statement of our approach to responsible AI, not a certification of a customer’s systems or a guarantee that AI outputs will be error-free.
2. Who we are and how to contact us
Nexa Point Group Ltd is registered in England and Wales under company number 16323341. Our registered office is 71–75 Shelton Street, Covent Garden, London WC2H 9JQ, United Kingdom.
For questions about this notice or a proposed AI service, contact info@nexapoint.co.uk. For privacy questions, personal-data rights requests or data protection complaints, contact privacy@nexapoint.co.uk. We will direct technical or governance concerns to the appropriate person for review.
3. What we mean by AI and automation
AI includes systems that generate, classify, summarise, predict, recommend or otherwise analyse information, such as machine-learning tools, generative AI assistants, AI-enabled cybersecurity features and software agents. Automation also includes rule-based workflows that do not necessarily use AI. Not every Nexa Point service, assessment or workflow uses AI.
Where AI plays a material role in a customer-facing activity, we seek to explain its purpose, the role of human reviewers, relevant limitations and any material use of third-party services through the applicable service description or engagement documentation.
4. Our responsible AI principles
- Lawful and purpose-limited use. We consider the intended purpose, legal basis, contractual authority and reasonable expectations of affected people before introducing AI into a workflow.
- Privacy and data minimisation. We seek to use only information necessary for an authorised purpose, apply appropriate access restrictions and avoid entering unnecessary personal or confidential information into AI tools.
- Human accountability. A named person or accountable team remains responsible for material customer decisions and actions supported by AI. Human review should be meaningful and proportionate to the risk.
- Accuracy and reliability. AI-generated content, classifications and recommendations can be incorrect, incomplete, biased or outdated. We consider appropriate verification and testing before relying on them.
- Fairness and accessibility. We consider whether an AI use case could disadvantage individuals or groups, including through biased data, inaccessible design or unsuitable assumptions.
- Security and resilience. We consider risks such as unauthorised access, data leakage, prompt injection, unsafe integrations, compromised suppliers and unexpected automated actions.
- Transparency and explainability. We aim to make the purpose and limitations of material AI use understandable and to provide an appropriate route to ask questions or request review.
- Ongoing oversight. We review relevant changes to models, integrations, data flows and risks, and may restrict, suspend or discontinue an AI-enabled activity when appropriate.
5. How we govern AI use cases
Our approach is risk-based. Before a material AI use case is introduced or substantially changed, the responsible team considers what the system will do, which information it will use, who could be affected, what a failure might cause and whether a simpler or non-AI solution would be more appropriate.
Depending on the use case and its risk, this may include documenting the business purpose and accountable owner; checking data sources and permissions; screening for a Data Protection Impact Assessment (DPIA) and completing one where required; assessing suppliers and contractual terms; testing outputs, security and fairness; defining human approval points; and setting arrangements for monitoring, incident response and withdrawal of the system.
Higher-impact uses, such as systems that could affect access to employment, finance, healthcare or other significant opportunities, require particular scrutiny and any safeguards mandated by applicable law. We do not treat an AI recommendation as a substitute for an authorised person’s professional judgement.
6. AI in our services and customer engagements
AI Governance and advisory. We may help organisations identify AI use cases, understand unapproved or “shadow AI” use, develop acceptable-use policies, maintain inventories and risk registers, assess supplier and data risks, and plan proportionate governance and staff awareness measures. Our assessments and advice support decision-making; they do not themselves constitute legal certification or regulatory approval.
AI and Automation. We may advise on or implement approved workflows, integrations, knowledge retrieval, productivity tools and AI-assisted processes. The scope of any AI agent’s access, actions, approvals and auditability is agreed for the relevant engagement. Where an automation could make consequential changes to a customer environment, we consider appropriate approval and rollback controls.
Microsoft 365 Copilot and cloud services. In an agreed readiness or deployment engagement, we may review identity and access settings, information permissions, data governance, licensing and relevant Microsoft 365, Azure, AWS or Google Cloud configurations. Use of a particular platform depends on the customer’s environment and agreed services; a technology partnership does not mean that every customer’s information is sent to that provider.
Managed IT Support, Cybersecurity and Infrastructure. AI-enabled tools may assist with identifying patterns, summarising incidents, categorising support requests or suggesting remediation. Recommendations require suitable validation. We do not treat a public enquiry or assessment as permission to access private systems, mailboxes, devices or networks.
Assessment Centre. Our public assessments may combine questionnaire answers with public technical indicators and rule-based scoring to highlight potential areas for improvement. An assessment score is advisory and does not establish that an organisation is secure, compliant or free from risk. More intrusive reviews require separately agreed scope and appropriate authorisation.
NexaAssist. NexaAssist is part of our product development direction. This notice does not represent that every proposed NexaAssist capability, customer portal or autonomous function is currently live. Any pilot or operational release involving additional personal-data processing will be subject to relevant documentation and controls before use.
7. Information that may be processed in AI-enabled work
Depending on the activity, information may include business contact details, service enquiries, assessment responses, documents or text supplied by an authorised user, approved knowledge-base content, support-ticket descriptions, configuration or diagnostic information, security alerts, prompts and generated outputs, user feedback, and technical logs needed to operate or secure a service. Some of this information may be personal data or confidential customer information.
We seek to limit information to what is necessary, reduce or remove identifying details where feasible, and use appropriate secure channels for sensitive material. We ask visitors not to submit passwords, multi-factor authentication codes, private keys, patient records or other unnecessary sensitive information through general website forms or public AI interfaces.
The relevant lawful basis, data protection role, retention criteria, recipient categories and individual rights are described in our Privacy Policy or the specific service privacy information. Where processing involves special-category or criminal-offence data, additional legal requirements apply. A new use of personal information for model training, testing or another purpose requires its own assessment and applicable notices, authorisations and lawful basis.
8. Third-party AI providers and technology partners
Some services may use technology supplied or hosted by other organisations, including cloud infrastructure, Microsoft 365 and Copilot-related services, AI model or API providers, security and monitoring products, integration platforms, document systems and specialist assessment tools. A named technology or vendor relationship does not mean that provider receives data from every Nexa Point service.
Before sharing customer information with an external AI service for a material use case, we consider the supplier’s role as controller or processor, permitted purposes, confidentiality, security measures, data retention, use of prompts and outputs, model-improvement or training terms, sub-processors, incident handling, location of processing and relevant contractual safeguards. Where required, we use appropriate data processing terms and obtain the necessary customer approvals.
Our standard approach is not to repurpose customer confidential information to train general-purpose AI models for Nexa Point’s own unrelated purposes without an appropriate lawful basis and authorisation. The separate terms of an approved third-party provider may govern how that provider handles prompts and outputs; relevant arrangements must be assessed for the actual service rather than assumed from this notice.
Where we act as a processor for a customer, we handle personal information within the documented instructions and agreed service terms. Where we determine the purposes of our own business operations, we may act as controller. The allocation of responsibilities depends on the processing activity.
9. International access and data transfers
Our UK operations may be supported by authorised personnel and service providers operating internationally. Depending on the relevant service and configuration, personal information may be stored or accessed outside the United Kingdom. UK hosting alone does not necessarily prevent international access.
Where a transfer to a separate organisation outside the UK is a restricted transfer under UK data protection law, we assess and apply an appropriate legal mechanism, such as applicable UK adequacy regulations or suitable contractual safeguards, together with any required transfer risk assessment and additional measures. Customer agreements may impose further restrictions. Further information about transfers relevant to a particular service can be requested at privacy@nexapoint.co.uk.
10. Human review, automated decisions and AI-generated outputs
AI may help organise information, suggest responses or identify possible risks. Such outputs may be uncertain or wrong and should not be relied upon without appropriate checks. The degree of review depends on the use case and potential impact.
Our general website assessments are intended to produce organisational recommendations, not to make decisions about an individual’s legal rights or similarly significant interests. If we introduce processing that involves solely automated decisions with legal or similarly significant effects on individuals, we will assess the applicable UK data protection rules and safeguards, including relevant information, opportunities to make representations, human intervention and the ability to contest the decision. Additional restrictions apply where special-category personal data is involved.
We may decline to deploy or continue an AI use case where the risks cannot be managed appropriately, where adequate authorisation is missing, or where the output is unsuitable for the proposed purpose.
11. Fairness, accessibility and responsible use
We consider the potential for discrimination, unequal outcomes and misleading outputs, particularly where AI affects people rather than only technical systems. Where relevant, we seek to assess the suitability of data, test performance across appropriate scenarios and provide a route for human challenge or correction.
We do not support using our services to gain unauthorised access to systems, unlawfully monitor individuals, misuse personal data or misrepresent AI-generated content as independently verified fact. Customers remain responsible for their own lawful use of tools and for any decisions they take on the basis of advice, subject to our agreed contractual responsibilities.
12. Security, monitoring and incident handling
We apply proportionate security and confidentiality arrangements to AI-related information and integrations, taking account of the service and risks. These may include access controls, appropriate authentication, segregation of environments, supplier due diligence, testing, logging, secure transmission and review of automated actions. No technical system can guarantee absolute security.
We investigate suspected misuse, inaccurate high-impact outputs, privacy incidents and security weaknesses through the applicable support and incident procedures. Where an incident involves personal information, we assess the notification obligations that apply to us and to the relevant customer or controller.
13. Assessments, enquiries and marketing
Where an assessment or AI-related enquiry allows visitors to request a consultation or further assistance, we use the information needed to respond and manage that request. Requesting a report or consultation does not, by itself, subscribe someone to promotional emails. Marketing is handled in accordance with our Privacy Policy and applicable electronic marketing rules. Any optional marketing choices are kept separate from necessary service communications.
14. Your rights, questions and complaints
Where we process your personal information as controller, you may have rights to access, correction, erasure, restriction, objection, portability and safeguards relating to significant solely automated decisions, subject to the applicable conditions and exceptions. You may also ask us to explain how AI has been used in a relevant process or request an appropriate human review.
Email privacy@nexapoint.co.uk to exercise a data protection right or raise a complaint. We will acknowledge a data protection complaint within 30 days, investigate appropriately, keep you informed and provide an outcome without undue delay. You may also complain to the Information Commissioner’s Office at ico.org.uk. Our Privacy Policy provides further information about rights, lawful bases and complaints.
15. Review and updates
We review this notice when material changes occur to our services, AI processing activities, legal obligations or governance approach. Changes will be reflected in the published version with an updated date. Service-specific contracts and notices may contain additional requirements that apply to a particular engagement. Please read this notice alongside our Privacy Policy, Cookie Policy and, where relevant, Recruitment Privacy Notice and the privacy or contractual information provided for a specific service.