Sustainability data: turning the tide Exhibit 4: Example ESG profile from the AllianzGI SusIE showing Company A in the insurance sector Image taken from SusIE dashboard showing ESG screening and controversy scores across several factors. As such, we put in place the Sustainability Methodologies – Generalist – these provide a full range offering of ESG and Analytics team in 2021 to develop a transparent, services, including raw data, peer analysis, ESG scores robust, and digital architecture. Now launched, this and qualitative analysis on ESG factors. architecture is known as SusIE or our Sustainability – Specialist – a highly granular, specific or niche offering of Insights Engine. data, opinions (eg, small cap, private markets) or themes SusIE will support the delivery of ESG, sustainability and (eg, biodiversity, social impact) impact data – with accompanying perspectives and – Technology – providers using alternative data capture opinions – to our investment professionals. It involves techniques like artificial intelligence or natural language a robust approach to scoping data sources, efficient processing to offer a new range of services and analytics. aggregation of this data, and the population of this These can include forward-looking measures, fuller data into clear and commercial front office-facing tools. coverage of investable universes, news flow screens and SusIE is designed to support the introduction of new truly independent raw data sets. data, removal of redundant data, and alignment to new Positioned for an ESG technology-powered future client product offerings. An example is the new net zero The one constant in any form of technology is change alignment toolkit, scheduled for launch later in 2023. and we are readying ourselves for this. We currently How we scope data for SusIE identify three likely important elements of any evolution of data and how they can be embedded in a robust and We value the cognitive diversity of the different resilient ESG data strategy: approaches in the market, and this helps us avoid 1. Evolving data capture away from current reliance on unintended implicit biases in model outputs. But purchased data. New technologies will allow for more discipline is needed in how we treat the different sources, alternative and independent sources, capturing more and we separate data providers into three broad perspectives. A likely future hybrid model will see quality 11 categories raw data complemented by external expert opinions. 3