AI & Digital Skills

Digital and AI tools are entering care services faster than governance can keep up. We train staff to use them well — and safely — covering digital care records, AI-assisted note writing, data protection boundaries and what must never be put into a public AI tool.

What’s involved

• Digital care records: recording well, not just recording
• What AI can and cannot safely do in a care setting
• AI-assisted note writing, summaries and rotas
• Choosing the right assistant: Copilot, ChatGPT, Claude, Gemini
• Enterprise vs free accounts — why the account type decides the risk
• UK data residency, retention and training-on-your-data settings
• Confidentiality, GDPR and prohibited AI inputs
• Recognising bias, error and hallucination
• Cyber security and phishing awareness for care teams
• Building an internal AI usage policy

Tools and models we train on

We are deliberately platform-neutral. Staff are trained on the assistants your service actually licences, and on the account settings that decide whether their use is safe.

• Microsoft 365 Copilot

  The safest starting point for most providers: it works inside your existing tenant, respects the permissions staff already have, and is covered by NHS-reviewed information governance guidance. Best for email, meeting notes, policies and reporting.

• ChatGPT (Team & Enterprise)

  Strongest general-purpose assistant for drafting, training material and analysis. We train on the business tiers only, where your content is excluded from model training and admin controls and retention settings can be evidenced.

• Anthropic Claude (Claude for Work)

  Excellent for long documents — policies, contracts, inspection reports and care plan reviews — and noticeably cautious in its answers, which suits regulated settings. Business plans exclude your data from training and support enterprise controls.

• Google Gemini for Workspace

  The natural fit for services already running Google Workspace, with the same admin, retention and data-region controls applied to Docs, Gmail and Meet summaries.

• Private deployments (Azure OpenAI, AWS Bedrock)

  Where personal or special category data must never leave a controlled environment, the same models can be run in a UK region under your own tenant with logging and retention you own. We cover when this is proportionate and when it is over-engineering.

• Open-weight models (Llama, Mistral) on your own infrastructure

  For providers wanting full data sovereignty or offline use. We set out realistic capability, cost and the governance burden you take on by hosting a model yourself.

• Assistants inside your care systems

  The AI features now built into digital care record and rostering platforms. We train staff on where these are genuinely safe to rely on and where a registered professional must still own the record.

What changes for you

• Time saved on administration and handovers
• Staff using digital tools confidently and compliantly
• A documented AI usage policy and governance position