Test, Trial & Evidence AI for Social Care

Evidencing the value of Artificial Intelligence (AI) to enable people to navigate, access and ‘have a conversation with’ social care information, advice, guidance and processes.

THE PROPOSAL

Working with a small group of local authorities, this collaboration will investigate the power of artificial intelligence to help people better engage with, and more easily navigate, social care services – whatever their situation or role. This work will be led by an established technical partner, who has developed and deployed AI services in both councils and the NHS.

This project will focus on ‘curated content’ specifically in Adult Social Care and/or Children’s Social Care. The project will not include any personal information or data, but test and evidence people’s engagement with social care advice and guidance, simplifying and allowing people to engage with it on their own terms – be that time of day, specific questions they would like to pose, the language they would like to communicate in as well as platform they choose to engage on.

Social care information and advice is especially hard to navigate, not only due to the sheer volume of information, the range of sources (local authority, NHS, third sector etc), static versus changing or time based services and content, but also due to the terminology often used, the complexity of processes and the limiting structure imposed by websites and forms.

Layer this on top of the average time someone spends on a website, that people only read 20-28% of what is on a webpage and the GOV.UK guidelines to write websites for a reading age of 9 years old – communicating social care information to people (who are often in crisis) becomes increasingly hard.

Imagine being able to give people unfettered access to ‘curated content’ – ie: content you know is correct, written by experts, either published and maintained by you or by trusted partners. Then allowing someone to work through a scenario based on their own circumstances, not that of a website structure or social care process. And then add on top of that the capability of proactively sending people links to local groups or services and personalising content based on an individual’s situation.

By laying an AI layer over validated curated content, we are able to do just that. People will be able to access the information they require much more easily and quickly. 

What we want to do is test, deploy and validate the approach, and quantify the benefits both to your service users and the council.

HEADLINE BUSINESS CASE

“At a time of ongoing financial challenges for many local governments, artificial intelligence presents the public sector with new opportunities to streamline services, free up staff time, enhance operations, and optimise resources.” (Govt Transformation)

Councils have hundreds – maybe thousands – of pages of content on their websites. Social care content is often the hardest to understand, as it has many links to third party content and organisations, and the way in which it is structured means that it can be out of date very quickly and/or complicated to navigate.

  • We want to allow people to have a conversation with your curated, up to date social care content;
  • We want people to ask the questions their way, using their language and preferred channels;
    • By language we mean in their own language (e.g: Polish or Urdu) as well as in the type of language they prefer – ie: more conversational and colloquial;
  • We want to use only the information that you – or trusted partners – already provide to people; 
  • We want to understand the art of the AI possible in terms of how it can transform social care and better manage demand and resources;
  • We want to quantify the time AI will save the front door;
  • We want to help manage the volume of assessments, referrals and applications;
  • We want to make accessing social care information easier

Critically much of the content this AI approach requires is already written and available either on your website, within wider social care materials or on the websites of trusted partners. By layering AI over this existing content, AI can lead a person through a conversation with that content, via popular channels – e.g. Whatsapp, Facebook Messenger, SMS or even as an app.

Whilst we are offering partners a choice in terms of the service users or content they would like to focus on, we expect the following groups and services to be of particular interest in terms of this project: 

  • Carers – ensuring better access to information, advice and local services
  • Reducing demand on the social care front door
  • Supporting people trying to engage with SEN services

BACKGROUND

“…the era of AI-driven local government has dawned.” – SOCITM

Over 30% of all local authorities are actively investigating AI. Projects dating back 5 years are enabling both internal and customer facing efficiencies. From transcribing social worker case notes, supporting recruitment processes, predicting needs and supporting services from waste to highways to FOI – as well as the widespread production of briefings – AI is being trialled and tested to support across all local authority services.

The social care front door is one of the most contact heavy, as many people still value a conversation. With demand increasing, the ability to deploy AI to support and enable people to have that conversation on their preferred channel, as well as ask questions and navigate social care services in a way that they understand and suits them, has the potential not only to be groundbreaking, but deliver huge time and resource savings.

“AI-powered assistants, technology, chatbots, and automated processes can allow better personalisation and micro-segment targeting. This will enable more efficient interactions, enquiry handling, and ensure people get access to the right services more effectively.”

“Bullet AI’s expertise helped us to create Leona, a virtual infant feeding specialist that’s made a real difference for our residents. We’re proud to recommend Bullet AI’s innovative approach.”

Medway Council Breastfeeding AI Approach “Leona”  – Commissioned by the Head of Health & Wellbeing Services, Scott Elliott

Medway Council Icon

Irrespective of the way that a resident interacts with Leona (i.e. via the website, app or WhatsApp), the information that she responds with is powered by a back office system that holds all the knowledge, advice and information that any infant feeding specialist would also learn during their induction period. The information that is held in the back office system will be quality assured by specialist infant feeding experts, so we can be confident that answers given by Leona, are in line with what we would expect any health professional to give. The back office system contains a bank of evidence based advice and when a resident asks a question, this information is interrogated and a suitable response is given to the parent. 

The system uses the latest machine learning technology and is programmed to ensure the questions are responded to in an empathetic, supportive and conversational manner. The ambition is that it leads to the parent receiving expert, tailored, rapid, evidence-based support at the touch of a few buttons. The additional benefits to technology of this kind is that it is fully multi-lingual. When Leona is asked a question in a particular language, the response is sent back in that same language. In addition, it can send voice notes and receive advice back in voice form – particularly useful if arms are full holding a baby.

TECHNOLOGY & INNOVATION

The underlying technology has already been built and is successfully working (see example of Medway council’s ‘Leona’ above). This existing AI technology and the learning gathered whilst building ‘Leona’ will support the collaborating councils move more quickly to successful deployment. 

The vast majority – if not all – of the curated content for your chosen use case will already be written, and the team will support the collation of all curated content to be accessed and queried. Sitting behind the AI technology is an existing data reporting capability which will give clear and comprehensive insight into how people are using the AI approach, what they are searching for, when, using what terms, what is returned to them and how the AI is learning.

With AI emerging rapidly and the use cases developing week on week, this project will also consider the ‘next steps’ in terms of people better navigating social care services supported by AI. One example – already in development, but not actively part of this project – is the ability to deploy an AI ‘form helper’ to support people as they complete application forms.

PROJECT DESCRIPTION

Councils will be working with a proven technical partner which will mean development and deployment will happen more quickly than relying on much sought after internal IT resources. The collaboration will be made up of a group of local authorities to share the wider knowledge, costs and benefits.

This collaboration allows your authority to drive forward a programme with technical experts and within 3 months have a live service focused on your chosen social care area or user group, which will then be live and refined for a further 9 months to assess the impact it has on the service.

  • Three to five councils together to investigate AI as a group;
  • Allowing each partner to decide on their chosen area within social care to test, deploy and evidence AI;
  • £850 worth of AI Model tokens (per partner), which are needed to deliver chat messages. Token usage is related to chat volumes and length of messages – we envisage this budget being more than enough for the project duration of 1 year;
  • Support of content curation & management;
  • Analysis of existing curated content and iterative expansion of curated content to ensure successful and useful deployment;
  • User group engagement & testing;
  • Ethics of content approach;
  • Management of risks;
  • Dashboards, data, analysis and in-depth reporting & learning;
  • Business case;
  • Shared knowledge, cost and learning

Each participating council will be able to access leading technical skills at a fraction of the market price, councils involved in the collaboration will help define the commercial model for the resulting ASC-AI technology and receive an ongoing benefit for helping evidence the ASC use cases.

WHAT ARE WE ASKING FOR?

  • £15,000 per partner 
  • 3-5 local authority partners

CONFIRMED COUNCILS

TIMEFRAMES

  • 12 month project 
  • Expected to run July 2024-June 2025
  • Actual date of project initiation will depend on all partners confirming

TECHNICAL PARTNER

Bullet Digital Solutions is a company based in the UK, specialising in web application development and digital transformation services. They help businesses improve efficiency by simplifying complex workflows and building customised web applications. Their team of experienced consultants has a strong track record of delivering solutions in various sectors, including the public sector, healthcare, and finance.

Bullet

GET IN TOUCH