POS Workplace Intelligence

AI and human oversight

Technology supports the work. POS remains responsible for the judgement.

Last updated: 9 September 2026 · Version 1.0

On this website

The form sends your enquiry to POS through its email service. It does not send the message to a generative AI model, score your organisation or make an automated decision about you. There is no public AI chatbot or live survey collection here.

AI-assisted work

AI-assisted tools support method development, permitted research, drafting and quality checks. Generated material can be incomplete, biased or wrong. It is not evidence simply because a model produced it. POS checks substantive claims against sources and reviews recommendations before client release.

Research must be suitable for the purpose and permitted for the intended use. External studies provide context; they do not replace evidence from the organisation being reviewed. AI-generated references, calculations and summaries require verification. Uncertainty and disagreement between sources must not be hidden to make a recommendation appear stronger.

Confidentiality first

The framework excludes raw employee responses, direct identifiers and confidential source documents from general-purpose AI prompts, embeddings and multi-client training collections. Public enquiries are not a training dataset. Removing a name alone does not make a document safe to share.

Any proposed AI processing of client evidence requires an approved purpose, minimum necessary inputs, agreed client data arrangements and a provider review covering access, retention, training use, subprocessors and transfers. An enterprise account or “no training” default is not, by itself, approval.

Human decisions

The review supports decisions about workplaces. It is not an individual employee rating system or an automated recruitment, disciplinary or redundancy tool. POS judgement remains distinct from evidence. Clients remain responsible for business and employment decisions, with appropriate professional advice.

Human review considers whether the evidence supports the conclusion, whether a small group or comment could identify someone, and whether the proposed action is proportionate. Technology does not replace these responsibilities or guarantee a particular business outcome.

Transparency and challenge

For commissioned work, the agreed method identifies material AI assistance and limitations. Ask how a finding was reached, what supports it and where judgement was used. Concerns about accuracy or fairness receive human review. Contact POS using the details below, identifying the finding or document concerned without sending raw employee responses through this public form.

We explain the relevant sources and reasoning, subject to confidentiality, third-party rights and the privacy of participants. Where a material error is identified, the finding and any affected recommendation require correction and reconsideration.

Accountability and scope

POS remains accountable for the work it supplies. This statement does not give permission for a new use of personal data or replace a client agreement, participant notice or provider assessment. Material changes to how AI is used require review of the data arrangements and appropriate information for those affected. The privacy notice explains how website enquiries are handled.

Contact POS

You can contact POS through the contact form on this site, by email at hello@posinteriors.com or by telephone on 01481 740222.

For privacy questions, rights requests or concerns about information shared, use these same contact details. Do not send identity documents or further sensitive information unless requested through an agreed route.