Charities harness AI for greater impact: Insights from the Deloitte Digital Connect panel session

As part of Deloitte Digital Connect, the Deloitte AI Institute and CAST recently hosted a panel session on the transformative potential of AI in the charity sector. Building on an introductory session to AI last month, the event featured speakers from Citizens Advice Scotland, HelpFirst, Catch22, CAST and the Deloitte AI Institute, sharing their experiences and insights on how AI is driving change and increasing the impact of their charitable initiatives. Below is a summary of the use cases presented at the event.

Use Cases

1. Citizens Advice Scotland: AI in crisis support

Speakers: Natasha Gilmour (Head of Operational Support) and George Holmes (Operations Manager)
Use Case: Extra Help Unit (EHU) AI Project

During the COVID-19 pandemic, the Citizens Advice Scotland’s Extra Help Unit (EHU) faced unprecedented demand, exacerbated by the subsequent cost of living and energy crises. The volume of consumers disconnecting from essential services surged by 1000%! It soon became apparent that traditional systems and increased staff numbers were insufficient to handle this surge. In response, Citizens Advice Scotland collaborated with HelpFirst and the Scottish Government’s CivTech accelerator to implement AI-driven solutions, including the HelpFirst Score and Risk Alerts, aimed at quickly identifying and prioritising support for the most vulnerable. This initiative significantly improved response times and alleviated staff stress.

Find out more (see slides 9–15)

2. Catch22: Optimising impact with AI

Speaker: Magid El-Amin (Director of Evidence and Insight)
Use Case: Predicting and Optimising Social Return on Investment (SROI)

Catch22 uses machine learning (ML) models to predict the Social Return on Investment (SROI) of their employability programmes. By inputting programme parameters into the ML model, they can calculate and forecast the wider value of participants returning to work, beyond the immediate value to the participants. This provides actionable insights for programme design and allows Catch22 to start continuous improvement processes early and deliver exceptional services from day one.

Find out more (see slides 16–21)

3. CAST: Experimenting with AI for digital support

Speaker: Dan Sutch (Director)
Use Case: Experimenting with a custom ChatGPT to scale digital support

CAST is exploring various AI applications to streamline workflows and enhance operational efficiency. Using their AI Experiment Canvas, their experiments aim to provide 20% efficiency gains. Their experiments include projects like “Sidekick” which aims to provide quicker digital support to charities at scale, as well as automating tasks such as note-taking, summarising data and drafting blogs. Another initiative is “Board Buddy,” a digital assistant designed to help charity trustees prepare for and participate in meetings. These AI tools aim to not only reduce the time spent on administrative tasks but also ensure that every charity can benefit from expert digital support.

CAST’s AI Experiment Canvas

Find out more (see slides 22–26)
View the AI Experiment Canvas
Find out more about CAST AI experiments

4. Deloitte AI Institute: PairD Platform

Speaker: Alexander Tan (AI Lead)
Use Case: PairD Generative AI Platform

Deloitte’s AI Institute has developed PairD, a secure generative AI platform designed to support 75,000 Deloitte users across various functions, including Tax, Accounting, Legal, and Audit. PairD offers tools for writing assistance, language translation, coding support, and more. Its backend design supports multiple large language models (LLMs) and integrates proprietary data, ensuring robust, cost-effective, and environmentally friendly operations. The platform exemplifies how large organisations can harness generative AI to drive innovation and efficiency while maintaining rigorous data security and compliance standards.

Find out more (sees slides 27–32)

Panel Discussion: Adoption and impact of AI

The panel discussion focused on key questions regarding the adoption and impact of AI within organisations, addressing practical challenges, and considering ethical implications. Below is a summary of the key discussion points.

1. Adoption of AI in organisations:

• Organisations from respective panelists are pushing forward with AI initiatives for a range of different reasons — to improve service delivery for Catch22, to enhance efficiency for CAST, and to protect staff and vulnerable users for Citizens Advice Scotland. AI also helps in identifying technical opportunities and solutions that were previously unknown.

2. Ethical and responsible use of AI

• There are significant discussions around the risks, ethics, and challenges associated with AI, especially for those without technical expertise. Key concerns include the potential for bias in AI models and the need for responsible data governance.
• AI should be used with clear boundaries and experimentation should be safe and transparent to ensure civil society’s influence in shaping the future of AI.

3. Addressing bias and risk:

• Organisations like Deloitte have developed their own internal AI tools to ensure data security and customisation specific to their needs, as existing market tools may not meet all their requirements. They’re keen to explore how we can democratise AI and make it accessible to everyone.
• It is crucial to address bias in AI models by ensuring the training data is cleansed and by being transparent about where the data comes from. Bias can be a significant challenge, and organisations need to collect comprehensive demographic data to effectively analyse and mitigate bias.
• Data governance and AI intersect in complex ways, and it is important to work through these issues with good advice and clear frameworks.

4. Practical tips for using AI:

• Start off by using AI to handle routine tasks so as to free up human resources for more complex and decision-making roles.
• For effective implementation, involve frontline users in the development process to ensure they buy into the AI tools and understand their value. This can be achieved through collaborative development and testing.

5. AI and job displacement concerns:

• There are concerns about AI taking away job opportunities, especially in creative fields. It is important to develop ethical policies that outline how AI will be used and to ensure it enhances rather than replaces human creativity and roles.
• In practice, AI tools should be used to support and enhance the work of human employees rather than replace them, focusing on tasks that can be automated to free up time for more critical work.

6. Storytelling and engagement:

• Bridging the gap between data-driven decisions and frontline anecdotal evidence can be achieved through storytelling. Engaging both data teams and frontline staff in the AI development process helps in creating tools that are both effective and trusted.

For more information on these initiatives and to explore potential collaborations, please contact me at david.scurr@wearecast.org.uk for a chat and to potentially connect you with respective speakers.

We’ve put together a living library of AI resources that we hope will support you as you explore the challenges and opportunities associated with AI. Please take a look — and do let us know if there is anything you need particular support with.