"With AI projects, you need a long breath"
RobecoSAM has been developing sustainability indices for large quoted companies for 20 years. In order to cope with the growing mountains of unstructured data, RobecoSAM implemented an AI solution together with ti&m. Manjit Jus, Managing Director, Head of ESG Ratings at RobecoSAM, talks about how the project went and what the companies learned about themselves during the process.
ti&m: What does RobecoSAM do?
Manjit Jus: Since 1999, RobecoSAM has run its annual Corporate Sustainability Assessment (CSA) which collects sustainability data on 4,700 companies and it will be 7,200 by March 2020. RobecoSAM collects data from some of the world’s largest companies directly through an online assessment platform as well as through publicly available information published on company websites, corporate annual reports and financial filings and sustainability reports. The CSA is widely recognized as the leading global assessment framework for sustainability ratings, and our data has been feeding the world-renowned Dow Jones Sustainability Indices since they were created in 1999. As of June 30, 2019, RobecoSAM managed USD 24.3 billion of client assets (including consulting and license agreements).
What is the difficulty with the evaluation?
The amount of sustainability data available has been increasing substantially over the last years, as a result of increasing global demand for transparent sustainability information by investors and other stakeholders. While global sustainability reporting standards exist, the type of reported information is constantly changing, in line with evolving global sustainability challenges. As a result, analysts analyzing this information are confronted with increasing data volumes, sources and sometimes incomparable data sets. Analyzing and processing this information is therefore becoming an increasingly complex and labor-intensive task.
How and why did you come up with the idea of using AI?
The increasing availability of large data sets allows for new technologies to be applied. We now have 20 years of data history – a perfect use case for applying technologies like AI and NLP. Increasing standardization within sustainability reporting has also allowed to apply these kind of technologies. Furthermore, new niche market entrants have helped support the adoption of these technologies in the sustainability space. It is a dynamic, rapidly changing space and traditional research frameworks benefit from leveraging new and flexible tools.
Why did you decide to go with ti&m?
We assessed a number of different partners, but we appreciated ti&m’s flexible, innovative and entrepreneurial approach – this fits well with the corporate culture at RobecoSAM. So far, the collaboration has been excellent, as the ti&m team members we have worked with are not only interested in the technologies, but also the content of the work we do. There is a clear passion emerging for sustainability, and this interest and excitement has made the collaboration feel more like a partnership rather than a one way business relationship.
How did your employees react to the AI project?
Everyone in the company who you talk to is excited about this, even if it doesn’t directly impact their areas of work. It shows strong management commitment, shows the company is staying innovative and a leader in its space, and sends a strong signal to our clients and corporate stakeholders that we are forward looking. Sustainability is an evolving theme, and being prepared to tackle tomorrow’s challenges is a key element. The people directly involved in the project are excited, as they are learning new things every day – regardless if they work in IT or in business.
What is your conclusion so far?
We learned a lot about ourselves and the way we work through this process. Going through a process like this requires a lot of preparation and hard work cleaning and reviewing data. By doing this, you discover a lot of new things about your own work – things that could be optimized or improved, processes that could be automated or replaced. Therefore, the learning experience goes far beyond just how these new technologies can positively impact your business, but it also forces you to make changes in your internal processes and systems to better prepare for a production- level roll out. Based on learnings we are not adjusting some key internal processes to be able to process data more effectively to feed future iterations of the models we have started building. This not only makes the AI implementation process smoother, but overall will improve the quality and effectiveness of our work.
What tips can you give to someone who also wants to try AI?
Have a clear vision of where you want to be, not tomorrow but in a year, in two years in three years. Find a partner that can support a realistic roll out of this vision. I think many managers believe that with enough money, AI can solve all of their problems in a short period of time. This simply isn’t true – while it’s about technology, it’s still about people and the organizational culture and how to support these. We see technology as an enabler, but in our field of work we also understand that analysts will continue to have a key role to play in interpreting the data we collect and make meaningful decisions based on this data. Therefore, longterm management buy-in is key. If necessary, break the project down into smaller tasks, spread budgets over a few years, but make sure to keep the momentum and excitement up internally. Make sure the project management is set up well, so that failures are turned into learnings and improvements.
Were you able to achieve your goals and to what extent?
This is to be seen. Our first goal is to improve the comprehensiveness and completeness of our assessment, by ensuring that all data is processed efficiently. We are not immediately trying to save time or money – we believe this will be a natural outcome of improving the way we work, freeing up time and resources to do other value-add work.