Danya Liu
Associate, Digital Industry
BloombergNEF
Two of the biggest challenges in improving a carbon footprint are determining its size in the first place – and analyzing the actions that might significantly impact it. The more data available, the more an individual or a corporation can make a plan and quantify the likely impact.
Over the past few years, the use of artificial intelligence has exploded, driven by cheaper computing and more available big-data sets. When applied to predictive maintenance, it can prolong the lives of industrial equipment and reduce O&M costs by 10%. Programmed into robots, drones, or cars, it can make possible object recognition and autonomous navigation. It already plays a part in sustainability because of its ability to reduce energy and resource consumption. Now there is a chance for AI to improve the way companies determine and report their sustainability impact.
The opportunity in corporate sustainability
Businesses of all sizes, sectors, and regions are stepping up their sustainability efforts in response to shareholder pressure. BNEF’s sustainability team tracks the pledges of companies committed to sustainability frameworks like the ones shown below. These commitments are a powerful lever in overall decarbonization. But behind the scenes, planning for corporate sustainability is a mammoth undertaking often mired in uncertainty. Uncertainty stems from not knowing total emissions volume, what extreme climate events will happen, when they will take place, and how they will impact existing assets. For investors, this uncertainty translates into risk.
Wide-scale collection and analysis of data using artificial intelligence might help meet the growing demand for corporate sustainability reporting. Startups have emerged leveraging AI technology to help track emissions accurately, to simplify reporting, and to enhance climate change strategy. Here we look at two such startups – Normative and Cervest. The former aims to bring sustainability data into external communications, while the latter targets internal decision-making.
Automating emissions reporting
One of the challenges in emissions reduction is taking stock of everything in the first place. This is especially true for calculating scope 3 emissions, essentially the footprint of all your suppliers and products in your value chain. Normative, a Swedish startup, has a software application that makes reporting easier. It goes as far as automating emissions reporting for scope 1 (direct emissions from operations), 2 (indirect emissions from electricity purchased) and 3 (indirect emissions from the value chain.
It does this by parsing through enterprise resource planning (ERP) data. Users upload procurement records from ERP systems, and Normative’s AI recognizes and pairs those products with the latest scientific research on linked environmental and social impact. Users can also upload energy bills, incident records, expense reports and various other enterprise datasets. The engine uses this to generate a bottom-up estimate of a company’s environmental and social footprint, along with recommendations on ways to reduce impact.
Normative’s service has attracted early customers with complicated supply chains, such as the building and construction industries. Other interested parties include private equity investors keen to standardize reporting across their portfolio companies.
Large companies can often spend hundreds of thousands of dollars hiring outside consultants to help them compile emissions and sustainability data. Normative says it can deliver the same service for a tenth of the price.
Modeling climate risk and impact
To help with internal decision-making, U.K.-based Cervest uses AI to model corporate climate risk. It offers a Geographic Information System (GIS) -based platform where different decision-makers can query the potential risk and impact that they care about. For instance, agricultural companies can model the impact of extreme droughts in Western Europe, or manufacturing companies can analyze which facilities are most at risk of weather-related disruption. The platform lets users model climate impacts over their chosen time-frame and also zoom in to specific sites.
To do this, Cervest uses AI algorithms to understand climate signals from satellite imagery and projections. Computer vision applications extract climate trends from remote sensing data, machine learning algorithms stitch all data sources into one integrated set, and finally statistical modeling predicts the impact of risks on assets.
By providing this model, Cervest hopes to illuminate the climate risks and implications for strategic decisions involving sites for expansion, or investments for climate resiliency, and even which insurance policies are worthwhile. So far its clients include research agencies, insurance companies, asset managers, as well as food and beverage companies.
BNEF Take
The power of AI lies in its ability to see patterns in large unstructured datasets, and predict machine, weather, and human behavior. Startups are particularly nimble at fine-tuning AI to suit specific industry applications. As climate and sustainability become an increasing concern for a range of industries, we would expect AI to be increasingly applied to these challenges. In addition to the two services profiled here, other companies are developing additional AI-based services in areas like building energy optimization and clean energy procurement.
These services aim to make the task of corporate sustainability more manageable and the reporting process easier. This in turn could encourage other corporates currently on the fence to commit to sustainability measures and take action.