Marketing Response Model Improvement using Behavioural AI
Project partners: Nationwide Building Society
Context: Nationwide Building Society is a British mutual financial institution, and the largest building society in the world with over 15 million members. It offers a wide variety of financial products.
Challenge: Most Marketing Response Models do not take into account the personal characteristics of individual customers; these characteristics are critical for Personalisation, which is known to drive success. Such information could be collected (e.g. personality questionnaire) but places a burden on the user. In order to remove that burden, there is a need for automatic extraction of linguistic and psychological characteristics from users from text samples. Scaled Insights’ Behavioural AI tools can do this!
Aims: To use natural language responses from a customer survey to predict responses to a selection of marketing campaigns so a predictive model can be built to drive Personalisation.
Project so far: Scaled Insights’ Behavioural AI was able to extract a variety of text and user
characteristics from anonymised language samples. The predictive models which used these characteristics significantly outperformed the non-text baseline model and doubled the correct identification of customers who responded positively to campaigns. This means that Scaled Insights’ segmentation model using linguistic and psychological characteristics derived by our Behavioural AI from customer panel text samples can increase the performance of Marketing Response Models.
Personalising Nudges for Motivating Learners in Healthcare
Project partners: Health Education England, University of Leeds
Context: Health Education England (HEE) is the national organisation responsible for the education, training and workforce development in the health sector. HEE’s Technology Enhanced Learning (TEL) Programme uses the most effective evidence informed technology and techniques to benefit health and care education. Their portfolio includes the design and development of e-learning resources for health and care professionals and the public.
Challenge: Health and care professionals at every level need to continuously develop and update their skills. This includes digital upskilling of the workforce that is necessary to realise the potential of state-of-the-art technologies like genomics and artificial intelligence in clinical practice. There is a need for effective and scalable e-learning solutions which caters to busy health and care professionals.
Aims: To use Scaled Insights Behavioural AI technology to automatically infer user characteristics from a natural language sample collected by asking a person to answer some open ended questions. Then to use these characteristics to personalise motivational messages (‘nudges’) to increase e-learning take-up.
Project so far: Scaled Insights has collaborated with academics from the School of Medicine and School of Computing at the University of Leeds to co-design two studies to investigate the feasibility of using Scaled Insight’s AI to infer user characteristics and to explore the relationships between these user characteristics and the effectiveness of different types of ‘nudges’ as motivational messages. To date we have we have established that the effectiveness of various types of ‘nudges’ differs between users with different characteristics that Scaled Insights’ AI has inferred. This means that Scaled Insights AI can be used for personalisation in the health education space.
Behavioural Insights into COVID-19
Project partners: West Yorkshire Combined Authority, University of Leeds
Context: The Covid-19 outbreak has had a huge impact on society. While certain aspects of everyday life were gradually returning to normal in the UK, local authorities needed to ensure the safe and healthy functioning of the society. West Yorkshire Combined Authority (WYCA) is responsible for ensuring economic prosperity supported by a modern, accessible transport network, housing and digital connections in the region.
Challenge: Citizens' behaviours during the Covid-19 outbreak are influenced by a complex set of factors, making the strategic delivery of the 'return to normal' a challenge for local authorities. More people-centric insights are needed to inform policy and city planning.
Aims: To better understand behaviours and attitudes of people in West Yorkshire as they relate to the Covid-19 outbreak, and thus enable WYCA to make more informed decisions about the next steps for West Yorkshire.
Results: Scaled Insights in collaboration with the University of Leeds conducted a large-scale online survey on attitudes and behaviours relating to COVID-19. WYCA has co-designed a set of questions to inform their policy decisions. Behavioural clustering based on Personality Traits obtained from language samples using Scaled Insights'
Behavioural AI technology has identified distinct groups of respondents whose actions and lifestyle behaviours during Covid-19 differed significantly. Innovative text analytics methods were also used to extract and categorise contextual factors influencing those actions.
This project assesses longer term impacts of COVID-19, and thus, future data collection is currently underway as a 3-month follow up, and is planned to continue in March 2021 to assess change overtime.
Behavioural Insights into groups identified as at high risk of COVID-19
Project partners: University of Leeds, UCL, University of Birmingham, US Military
Context: Due to the unprecedented and rapidly changing impact of the COVID-19 outbreak, it is imperative to understand and appropriately support people who have and continue to manage health conditions. Responses to the outbreak have led to varied actions across the world to reduce infection and spread.
Challenge: Help the UK Government, public health authorities and charities to understand how to communicate with people about protecting themselves from COVID
Aims: To understand peoples' thoughts and behaviours relating to the coronavirus (COVID-19) outbreak, and how this has specifically impacted people identified as ‘vulnerable’.
Results: In collaboration with our partner organizations, Scaled Insights conducted a large-scale online survey to assess awareness, attitudes and actions of UK adults identified as at risk of severe illness from COVID-19. Behavioural clustering based on Personality Traits obtained from language samples using Scaled Insights' Behavioural AI technology has identified distinct groups of respondents whose wellbeing, depression, actions to manage their safety and health conditions during COVID-19 differed significantly. Our partners and the UK Government are using these insights to tune behaviour change and safety messaging to both specific groups as well as the public at large. We have been invited to submit a paper for a prestigious health journal discussing our findings.
This project was extended by our Partners and will now also assess longer term impacts of COVID-19 on these populations, making this an ongoing project with future data collection will commence in September 2020 and March 2021 to assess change over time.