Research, guidance and local context

Evidence and research behind BlueCap AI

BlueCap AI uses external research, official guidance, academic studies, sector evidence and local strategy to shape how we design practical AI support.

Using the evidence responsibly

How to read this page

The sources below serve different purposes. Some describe national or local need. Some provide practical guidance. Others examine how people, staff teams or organisations are using AI.

They help explain the needs, risks and design choices behind BlueCap AI’s approach. They do not prove that BlueCap AI has achieved a particular result.

Our own outcomes are collected and reported separately using delivery evidence such as practical outputs, participant feedback, confidence change, follow-up needs, partner feedback and anonymised examples where consent is clear.

References to local or regional strategies show alignment and context unless a confirmed relationship is explicitly stated.

Warrington Borough Council supports and endorses BlueCap AI’s local work and is helping explore funding routes for future sessions. Any specific funded or commissioned programme will be described separately once formally agreed.

Research, policy and guidance can change. We review this page periodically and link to the original source so readers can check the latest version.

Last reviewed: 2 August 2026

Practical AI learning and implementation

Access to an AI tool is not the same as knowing how to use it safely or usefully.

Current UK evidence suggests that many people need support with understanding risks, protecting information, checking accuracy and applying AI to relevant tasks.

Organisations may also adopt AI tools without embedding them into clear workflows, staff practice or governance.

BlueCap AI therefore focuses on practical tasks, clear boundaries, supported practice and one useful starting point at a time.

Source 1

AI Skills for Life and Work: Summary report

Department for Science, Innovation and Technology, Department for Culture, Media and Sport, and Ipsos 2026

UK research found that confidence with AI remained limited and identified information security, accuracy checking and risk awareness as important AI skills.

The research found that 17% of respondents said they could explain AI in detail, 28% felt confident using AI in daily life and 21% felt confident using it at work.

It also identifies protecting information, judging accuracy, recognising AI-generated material and understanding risks as important AI skills.

How this informs our approach

BlueCap AI includes privacy reminders, output checking and practical confidence-building rather than assuming that access to AI means someone is ready to use it independently.

Read the AI Skills for Life and Work report

Source 2

Skills for AI: What Works for AI Upskilling in the UK

Skills England and Royal Holloway, University of London 2026

The report recommends practical, accessible and task-based AI learning supported by coaching, clear language and opportunities to practise.

The report recommends AI learning that is practical, reachable, modular, personalised and connected to real tasks.

It highlights plain language, mobile-friendly options, low-technology alternatives, coaching, safe spaces to practise and ongoing support.

How this informs our approach

This closely matches BlueCap AI’s task-based and human-supported coaching model, including paced learning, follow-up and accessible materials.

Important context

The evidence covers workplace training alongside informal, community-based, voluntary and self-directed learning. Its Good Things Foundation case study examines entry-level AI literacy for people at risk of digital exclusion, including plain English, everyday tasks, bite-sized learning and guided support.

This provides strong design guidance for BlueCap AI’s approach. It does not prove BlueCap AI’s own outcomes.

Read the Skills for AI reportExplore the supporting case studies

Source 3

Artificial intelligence in UK businesses: 2023 to 2026

Office for National Statistics 2026

AI use is increasing across UK businesses, but adoption often remains limited in depth.

Among UK businesses with at least ten employees, reported use of one or more AI technologies increased from around 12% in late 2023 to around 35% in June 2026.

The evidence also suggests that adoption often remains relatively shallow.

How this informs our approach

BlueCap AI distinguishes between experimenting with AI and implementing it through suitable workflows, staff support, governance and evidence.

Important context

Reported adoption does not show that the technology is being used effectively, safely or consistently.

Read the ONS analysis

Digital inclusion and public-service access

Digital inclusion is not only about whether someone owns a phone, tablet or computer.

People may also face barriers involving confidence, skills, accessibility, connectivity, trust, service design and a lack of suitable human support.

Digital services should not remove non-digital routes for people who need them.

BlueCap AI uses calm coaching, trusted community settings, plain language and options such as voice, dictation and read-aloud support where these help the individual.

Source 4

Digital Inclusion Action Plan: One Year On

UK Government 2026

The national plan recognises that digital inclusion requires coordinated work across access, skills, confidence, local support and service design.

The national action plan recognises that digital inclusion requires coordinated work across access, skills, confidence, local support and service design.

It emphasises the importance of government, local organisations, charities, communities and businesses working together.

How this informs our approach

It supports locally delivered and human-supported digital inclusion rather than treating device access as the complete solution.

Read the Digital Inclusion Action Plan update

Source 5

Inclusive digital healthcare: a framework for NHS action on digital inclusion

NHS England 2023

NHS England states that digital healthcare should be inclusive and should remain complementary to non-digital support.

The framework states that digital healthcare should be designed and implemented inclusively.

It also makes clear that digital approaches should remain complementary to non-digital support.

How this informs our approach

It supports practical help with NHS App confidence, accessibility and service navigation while keeping human and non-digital routes available.

Important context

BlueCap AI supports confidence, preparation and navigation. It is not a clinical service and does not provide medical advice.

Read the NHS England framework

Source 6

Ageing Well in Warrington: Joint Strategic Needs Assessment

Warrington Borough Council 2025

Local evidence shows that older Warrington residents face significant barriers involving digital access, skills and use of online health services.

The assessment reports that 20.3% of Warrington residents aged 65 and over are digitally excluded and 25.6% have limited or no digital skills.

It also reports that daily internet use falls to 49% among residents aged 85 and over, and that older residents are less likely to use several GP online services.

How this informs our approach

This supports accessible, patient and human-supported digital inclusion work for older adults in Warrington, including practical help with confidence and service navigation.

Important context

These are population-level findings based partly on 2023 data. They demonstrate local need, not demand for BlueCap AI specifically or evidence of BlueCap AI’s impact.

Read the Warrington ageing assessment

Source 7

Quality Strategy for NHS-funded care in England

NHS England 2026

The latest NHS direction combines expansion of digital and AI-supported services with a continued commitment to non-digital routes and in-person support.

The strategy says the NHS App must complement non-digital services and notes that libraries are providing in-person support with the App.

It says support must remain available for digitally excluded communities and sets an expectation that AI will help users navigate selected NHS App pathways by the end of 2026/27.

How this informs our approach

This strengthens the case for accessible, human-supported NHS App confidence and navigation work, while retaining suitable non-digital support.

Important context

This is policy direction, not evidence that planned AI features are effective. BlueCap AI supports confidence and navigation; it does not provide medical guidance.

Read the NHS England Quality Strategy

Unpaid carers and local need

Unpaid carers may manage appointments, medication, forms, messages and service coordination alongside their wider caring responsibilities.

Research and local strategy show the need for clear information, accessible support and more joined-up services.

AI may help someone prepare, organise, draft or understand information. It should not replace professional advice, care decisions or safeguarding support.

Source 8

Unpaid Care Dashboard

Centre for Care 2025

The dashboard uses Census evidence to show the scale and local distribution of unpaid care in England and Wales.

The dashboard reports that the 2021 Census identified approximately five million unpaid carers in England and Wales, providing care valued at £162 billion.

It also allows local-authority-level analysis.

How this informs our approach

It provides a reliable basis for understanding local need and prevents us from relying on vague or guessed statistics.

Important context

Census evidence describes the scale of unpaid care. It does not show how many carers need AI support or whether a particular intervention is effective.

Explore the Unpaid Care Dashboard

Source 9

A fresh approach to supporting unpaid carers

Carers UK 2025

Carers UK research highlights the substantial administrative workload many unpaid carers manage around NHS services.

Carers UK reported that 34% of carers surveyed spent at least ten hours per month on NHS administration for the person they cared for.

Eighty-four per cent said more joined-up services would reduce the time spent arranging and managing NHS care.

How this informs our approach

It supports practical help with preparation, organisation, appointments and communication while keeping decisions and formal advice with appropriate professionals.

Important context

The research shows the scale of administrative pressure. It does not show that AI or BlueCap AI automatically reduces that burden.

Read the Carers UK report

Source 10

Warrington Carers Strategy 2025–2028

Warrington Borough Council and local partners 2025

The strategy commits local partners to accessible, responsive and joined-up support for unpaid carers.

The strategy sets commitments to ensure unpaid carers are listened to, valued and supported through accessible and responsive services.

How this informs our approach

It supports BlueCap AI’s focus on practical, accessible help for unpaid carers with preparation, organisation, communication and safe use of digital tools.

Local partnership context

Warrington Borough Council supports and endorses BlueCap AI’s local work and is helping explore funding routes for future sessions.

Any specific funded or commissioned programme will be described separately once formally agreed.

Read the Warrington Carers Strategy

Responsible AI, privacy and human oversight

AI should support people and accountable staff.

It should not replace professional judgement, safeguarding or important decisions.

Responsible use requires a clear purpose, suitable tools, minimal necessary information, transparency, named accountability, human review and routes for feedback or escalation.

Human review can reduce risk, but it does not automatically make every AI workflow safe.

Source 11

Data and AI Ethics Framework

UK Government 2025

The framework focuses on privacy, fairness, transparency and protecting people and communities from harm.

The framework focuses on privacy, fairness and protecting individuals, communities and society from harm.

It also supports documenting ethical decisions throughout a project.

How this informs our approach

BlueCap AI builds clear purpose, risk controls, evidence capture and decisions to continue, change or stop into pilot design.

Important context

This is guidance rather than certification. It does not replace legal, safeguarding, professional or local governance requirements.

Read the Data and AI Ethics Framework

Source 12

AI Playbook for the UK Government

Government Digital Service 2025

The playbook provides practical guidance on AI capabilities, limitations, risks, testing, deployment and secure use.

The playbook provides practical guidance on AI capabilities, limitations, risks, selection, testing, deployment and secure use.

How this informs our approach

It supports starting with a defined need, checking approved tools, testing before wider use and maintaining human oversight.

Important context

BlueCap AI readiness support helps organisations identify practical questions, risks and next steps. It is not a formal compliance or assurance audit.

Read the AI Playbook

Source 13

Guidance on AI and data protection

Information Commissioner’s Office Current guidance

The ICO guidance covers fairness, transparency, data minimisation, accountability, security and individual rights.

The guidance covers lawfulness, fairness, transparency, data minimisation, accountability, security and individual rights where AI processes personal information.

How this informs our approach

BlueCap AI avoids unnecessary personal identifiers, includes privacy reminders and flags workflows that need proper organisational review.

Important context

Data-protection law and guidance can change. BlueCap AI does not provide legal or formal data-protection advice.

Read the ICO guidance

Source 14

AI in Care Alliance

Digital Care Hub and partners 2026

The Alliance places dignity, choice, control, equality, wellbeing and co-production at the centre of responsible AI in care.

The Alliance places human rights, independence, choice, control, dignity, equality and wellbeing at the centre of responsible AI use in care.

It also emphasises co-production with people drawing on care, unpaid carers, care workers and providers.

How this informs our approach

BlueCap AI aims to involve the people affected by a service in shaping, testing and reviewing how AI is used.

Learn about the AI in Care Alliance

Source 15

Scribe and prejudice? Exploring the use of AI transcription tools in social care

Ada Lovelace Institute 2026

Research across 17 local authorities found potential benefits from AI transcription alongside risks involving errors, safeguards and accountability.

This qualitative research across 17 local authorities found reported benefits from AI transcription tools.

It also identified risks involving inaccurate records, hallucinations, inconsistent safeguards and unclear accountability.

The research found limited evidence about the experiences of people drawing on care.

How this informs our approach

It supports beginning with lower-risk use cases, evaluating more than time savings and recording errors or failures as well as positive results.

Important context

The study focuses on transcription tools in social care. Its findings should not be treated as applying automatically to every type of AI use.

Read the Ada Lovelace Institute research

Accessibility and neurodivergent use

Accessible AI support should not assume that everyone learns, communicates or processes information in the same way.

BlueCap AI uses short instructions, predictable steps, plain language and one manageable task at a time.

Voice, read-aloud and dictation can help some people, but they may create barriers for others. We therefore aim to offer choices rather than one assumed format.

Source 16

Making Content Usable for People with Cognitive and Learning Disabilities

World Wide Web Consortium 2021

The guidance recommends clear purpose, understandable content, predictable navigation, personalisation and user involvement.

The guidance recommends clear purpose, familiar patterns, understandable content, predictable navigation, personalisation and involving people with cognitive and learning disabilities in design and testing.

How this informs our approach

It supports plain language, predictable flows, easy recovery from confusion and testing materials with the people expected to use them.

Important context

This guidance supports accessible design but does not replace formal accessibility testing or WCAG requirements.

Read the W3C guidance

Source 17

Exploring Large Language Models Through a Neurodivergent Lens

Proceedings of the ACM on Human-Computer Interaction 2025

The study found that neurodivergent communities use language models for communication, learning and productivity while also raising concerns about reliance and neurotypical assumptions.

The study analysed discussions across 61 neurodivergent online communities.

It identified uses involving communication, learning and productivity, alongside concerns about overreliance, neurotypical assumptions and text-only limitations.

How this informs our approach

It supports personalisation, multiple interaction options and clear boundaries around reliance on AI.

Important context

The study analyses online discussions and should not be treated as representing every neurodivergent person.

Read the study

Source 18

Envisioning Large Language Model Use by Autistic Workers for Communication Assistance

CHI Conference on Human Factors in Computing Systems 2024

The study explored possible benefits and risks when autistic adults used a language model for workplace communication support.

The study explored how autistic adults might use a language model for workplace communication.

Participants valued structure, privacy and availability.

The research also identified risks from ungrounded assumptions and neurotypical framing.

How this informs our approach

BlueCap AI uses AI to help people draft, prepare and explore alternatives rather than presenting one socially “correct” answer.

Important context

This was a small exploratory study and should not be treated as representing all autistic adults or workplace situations.

Read the study

Warrington and regional alignment

BlueCap AI is based around Warrington and works within a wider regional focus on digital inclusion, workforce confidence, responsible AI and support for unpaid carers.

The strategies below help explain the local and regional context for our work.

Source 19

Cheshire and Merseyside AI and Automation Sub-Strategy

NHS Cheshire and Merseyside Integrated Care Board 2025

The regional strategy calls for responsible adoption, clear governance, agreed use cases, workforce skills, public involvement and evidence before scale.

The strategy calls for responsible and ethical adoption, clear governance, agreed use cases, workforce skills, patient and carer involvement, evaluation and regular review.

How this informs our approach

It closely aligns with BlueCap AI’s focus on readiness, staff capability, public involvement and evidence before scaling.

Important context

This is regional strategic alignment. It does not itself represent an endorsement, commission or partnership with BlueCap AI.

Read the Cheshire and Merseyside strategy

Source 20

Warrington Corporate Strategy 2025–2029

Warrington Borough Council 2025

The council strategy includes reducing inequalities, supporting vulnerable residents and bridging the digital divide.

The strategy includes reducing inequalities, supporting vulnerable residents and bridging the digital divide through access to digital services and appropriate skills.

How this informs our approach

It provides a current local context for BlueCap AI’s community-based digital inclusion and practical AI support.

Local partnership context

Warrington Borough Council supports and endorses BlueCap AI’s local work and is helping explore funding routes for future sessions.

Any specific funded or commissioned programme will be described separately once formally agreed.

Read the Warrington Corporate Strategy

How we use evidence

External research helps us understand the needs, risks and design choices behind our work. BlueCap AI’s own outcomes are collected and reported separately using evidence such as:

  • Attendance and repeat engagement
  • An assistant or workflow created
  • A practical task completed or supported
  • Confidence change
  • Follow-up or signposting needs
  • Partner feedback
  • Anonymised examples and quotations where consent is clear
  • Barriers, errors and things that did not work

Our evidence language

We use cautious wording such as “research suggests” and “early evidence suggests” unless stronger claims are supported by direct evidence.

Clear boundaries

Our boundaries

AI can help people prepare, organise, draft, practise and understand information.

It should not replace:

Continue exploring

Choose the next useful route

Learn more about our method, use free practical resources or explore support for your organisation.

Back to BlueCap AI