In the final installment of this three-part series, I will zoom out and shift focus to ESG (Environmental, Societal, and Governance standards that measure an organization’s impact on society, the environment, and how transparent and accountable it is). Whilst focusing on Frugal AI is important, making it actionable requires a framework to support your organization through this journey.
Thankfully the ESG framework already exists in most organizations and, coincidentally, is aligned to Frugal AI objectives. When setting out ESG aspirations, the bigger challenge that organizations face is grappling with ESG as a board level priority, and getting the associated reporting ‘right.’ Can advanced analytics and AI alleviate these reporting pains? To start, let’s take a step back and talk about the evolution of ESG.
The Evolution of ESG
In record time, ESG has swiftly transitioned from a ‘side of desk initiative’ (tasks that need to get done, but goes over and above what a core job entails), to being equally important as digital transformation and a core element of an organization’s strategy. This shift in focus for organizations can be attributed to regulation and, on a global scale, the growing NGO-driven pressure to act ‘responsibly’ and:
- Be accountable for the effects of direct emissions on the environment attributed to their business, as well as the direct and indirect effects of their supply chain on the physical surroundings (Environment).
- Continuously assess and improve how they interact with the communities they operate in (e.g., consumers, suppliers, employees), how they contribute to wider goals that benefit humanity, and how they navigate ever-changing social movements (e.g., #metoo, #genderpaygap) (Society).
- Rethink, proactively govern, and monitor the distribution of roles and responsibilities to ensure that there are appropriate counterpowers internally; this could also include corporate policies, ethics and diversity, and company culture (Governance).
The short answer? Yes. In recent years, there have been proactive and reactive examples of ESG in action. Examples at Walmart and Volkswagen in the U.S. and Orpea in France demonstrate that having robust ESG strategy and implementation frameworks is a must have. Proactive action on ESG is a necessity and benefits such as averting negative impacts on the environment and positively shaping society outweigh the lure of inaction. Some organizations take it a step further and leverage ESG as a strategic asset; an increasing necessity to access funding. According to McKinsey, organizations such as Unilever are increasingly using ESG strategy and execution to their competitive advantage.
This value can be seen in (i) top line growth, (ii) cost reduction, (iii) reduced regulatory and legal interventions, and (iv) employee productivity uplifts. This is good news and a good incentive for companies aiming to leverage ESG as a strategic asset. Aspirations are great, but the challenge then is on execution.
The Case for AI and Advanced Analytics to Enable ESG Outcomes
As organizations increasingly implement ESG strategies into their business practices, there needs to be robust mechanisms to properly analyze and monitor the relationship between ESG and financial performance and outcomes cited in the aforementioned McKinsey article.
It is also no longer enough for firms to ‘say’ they abide by ESG, they have to prove it — to auditors, NGOs, and other stakeholders, and to do that they need data. Technologies such as AI and advanced analytics can be critical for companies to measure, achieve, and provide evidence of ESG activities objectives and reporting. However, a key criteria for successfully gaining value from advanced analytics and AI is collecting data to measure and baseline the current state. This raises the question, “Is it even possible to collect an accurate, consistent, dataset on the E, S, and G in ESG?”
In a recent study by Accenture Research, only 26% of organizations indicated that they had clear, reliable data to underpin their ESG KPIs. To help determine why that is, let’s first assess an ESG dataset in its respective siloed pillars.
On their own, data points for measuring environmental impact such as energy and water use, waste and pollution management, biodiversity, and land use data, can be plentiful, but collecting the data is complicated as there are few norms for organizations to abide by. Baselining and interpreting societal data is problematic as it is less quantitative, less standardized, and relatively subjective.
Governance data is even trickier. The definition and conventional wisdom of ‘good’ governance has shifted over time and has gone from balancing the needs of stakeholders to an emphasis of maximizing shareholder value, and now back to an emphasis on stakeholders. For both societal and governance pillars in ESG, how do you baseline and model scenarios such as bad labor practices, the impact of the recently reported U.K. gender pay gap, and if diversity on executive boards is a lever that contributes to maximizing shareholder value?
It is safe to conclude that, when viewed in silos, a robust dataset could be collected on some aspects of ESG, but not all. This is supported by the finding from Oracle that 91% of business leaders are currently facing major challenges in making progress on sustainability and ESG initiatives. Challenges include finding the right data to track progress and time-consuming manual processes to report on ESG metrics. Much of the data that needs to be collected either comes from many systems (i.e., internal to an organization as well as data from suppliers and suppliers of the suppliers, doesn’t exist yet, or is subjective). When you overlay starkly inconsistent ESG frameworks, and the inherent complexity of making sense of the data, perhaps advanced analytics and AI solutions can enable organizations to execute on ESG.
One way to make progress on ESG data collection and baselining is to embed ESG reporting into core operational and management systems and gradually automate processes and reporting. Analytics can then take reporting across the finish line. On the one hand, this approach will expose flaws and inconsistencies, but on the other knowing your position and what you ‘don’t know’ counts towards actionable goals, shifting the dial towards consistent data and most importantly, transparency and trust. More importantly, Analytics and AI have an instrumental role to play in taking action: Measuring is only a first step towards making change, which will demand full reinventing of critical processes, improving audits, and reinventing business models — all areas in which AI and analytics can play a strong part.
So, Where Do We Go From Here?
By now you could ask yourself: Wait… didn’t this start as a blog post series on Frugal AI? We seem to have gone far from the initial topic. And yet, not that far: If you have read my two first blog posts, Frugal AI is about making choices and not over consuming data and power compute, which should be one of the driving pillars for any company willing to align its AI ambitions with its ESG principles. Of course, making sure AI efforts are actually tuned towards ESG activities should be another driving principle embedded in AI strategies and monitored as part of well-designed AI Governance.
My call for action is simple: AI should support your ESG efforts, and should be done in a frugal, ESG-aligned manner. I challenge you to have a frank assessment of where you stand with aligning your roadmap on your company’s ambition.