How AI could aid ESG teams

With so much talk about the potential for artificial intelligence to transform the way we do things, it is perhaps no surprise that ESG and sustainability are squarely in the sights of those looking to capitalise on the latest innovations. It was noted at last year’s World Economic Forum meeting in Davos that without the power of AI the world will struggle to meet ESG goals and address climate change, and in private markets some ESG professionals are increasingly reaching the same conclusion.

“The issue for heads of ESG is the breadth and depth of what they are tasked with looking after,” says Victoria Gillespie, head of ESG at Alter Domus. “They really have a lack of time and they are inundated with information across a vast subject and typically with a global remit. We see AI as being able to help them get back some time, taking on some of the more mundane tasks to free them up for more strategic thinking, as well as enhancing their day-to-day processes by leveraging analytics.”

By freeing up ESG experts to spend more time on strategic initiatives, AI could dramatically increase the impact of efforts by private funds.

“What we are all really thinking about is how do you make the job of ESG integration and the role of the ESG professional easier by using AI to be more effective,” says Adam Heltzer, a partner and head of ESG at Ares Management. “The ESG task has multiplied in scale in recent years, so which parts of that are most important for a person to do and where can AI tools take over.”

Today, firms are just at the beginning of the journey: “Across the private markets industry, firms are exploring, mapping and prioritising new AI technologies and analysing their implications across activities,” Heltzer says. “ESG applications are in that queue but we are not at the point yet where everyone is going full force prioritising developing these tools.”

The first stage of AI adoption in ESG came with tools that could cross reference a private markets portfolio with thousands of global news sources on a rolling basis to keep GPs informed of any adverse stories. Those search engines were taught about what was relevant and what wasn’t, but the next stage could be much more transformative.

Heltzer says: “There is a huge focus today on engagement with portfolio companies and helping management teams understand what is most important on the menu of ESG items. For example, if we can automate and scale ESG data collection, we can free people to spend much more time on value-added work with portfolio companies to accelerate impact. The idea that you could bring some scale and leverage to the way you collect data, clean data and present data frees up the ESG professional to analyse the data and make strategic use of it in conversations with portfolio companies.”

Last year, Alison Humphrey, who was previously a senior ESG and sustainability executive at TPG and an adviser focusing on sustainable growth at investment bank PJT Partners, launched Arna, a capital advisory group focused on climate. She says: “From a reporting standpoint, one of the most exciting uses is employing generative AI to fill in CSRD [Corporate Sustainability Reporting Directive]-type regulatory compliance and reporting schemes. There is a lot of capital flowing into some of these new ESG data technology platforms, which think they can now auto-populate 60 to 70 percent of those reporting requirements already and will eventually get to 99 percent. That delivers significant capacity expansion to ESG teams that can be used to work on more strategic value-creation efforts.”

Humphrey also sees potential in AI’s ability to expand the capacity to analyse value-creation opportunities: “We are still in the phase of gathering the data for these massive language models, with great potential for an enhanced ability to test and simulate potential value-creation pathways. For now, it’s early days.”

Meanwhile, Gillespie points to AI’s role in risk management processes, where it could be used to look at regulatory risk within portfolios, for example. “We know that regulation around ESG has grown exponentially in the past decade, so if you can use AI to look at the materiality of what that regulation means for an organisation or a portfolio, that is a powerful tool.”

Augmenting ESG activities

This is not about saving headcount – those at the frontline say there is no effort to cut staff numbers but rather to advance ESG roles away from data input towards more impactful tasks. In the current learning phase, there is also a lot of oversight required to check on AI tools being used.

“The key challenge with AI as it relates to ESG data is the risk of ‘garbage in, garbage out’”

Adam Heltzer
Ares Management

“ESG professionals are meant to be engaging with portfolio companies to help them create well-tailored and resourced ESG roadmaps,” says Heltzer. “If you have a portfolio of corporate investments across private equity and private credit, as we do, you may be able to imagine a bot that can respond to the most common ESG questions, tailored to each sector. Then you can have the professional pull those pieces together with the CEO into a strategic roadmap. It is hard to replace humans in that strategic change management role but there is a lot that AI can do to lay the groundwork.”

Right now, there is a lot of work being done by service providers and consulting firms to develop AI tools that will work for private markets ESG teams. Heltzer says: “In the emerging space of AI, like many others under the ESG umbrella, we as target consumers are often training the service providers as to what we need. There is probably a period of at least a year or two now where we will be helping them refine their offerings before we hit widespread adoption.”

Private markets firms are at varying stages of their exploration of AI tools. “We see some firms that are quite new to this, various CFOs that are starting to use ChatGPT, and others that have set up whole departments tasked with AI adoption and creating proprietary tools, as we have at Alter Domus,” says Gillespie. “So, the extent to which AI is being used already really varies, and there are still lots of people taking a wait-and-see approach. Certainly, there is a common consensus around overwhelm within ESG, but a varied response.

“Some of the smaller PE firms are looking to outsource, just as they outsource their ESG strategy and implementation. They want to rely on existing service providers to use AI for their benefit, while we do see some building their own proprietary tools.”

Gap in understanding

At this stage, there is also a lot of concern about the risks of using AI. “I would say the biggest risk from AI is a lack of understanding,” says Gillespie. “There is a huge lag in terms of what people think AI is and what it can actually do, so the challenge is making sure you have staff capable of keeping pace with developments.”

“Adoption by GPs has been a bit slower than it might have been because of data sensitivities”

Alison Humphrey
Arna

Ares’ Heltzer sees another issue: “The key challenge with AI as it relates to ESG data is the risk of ‘garbage in, garbage out’. Can we really trust the AI tools to give us insights if the information going in isn’t clean and reliable and hasn’t been audited in the right way? There is perhaps a bit of a J-curve issue, where AI maybe amplifies those data problems where we stand today, but going forward it will be a hugely useful and impactful tool for saving time and delivering actionable insights.”

Arna’s Humphrey argues that one of the biggest barriers to AI use today is concern about data security. “The adoption by GPs has been a bit slower than it might have been because of data sensitivities,” she says.

“Most of these private equity and private credit firms are extremely concerned about their data in general. ESG and sustainability data is particularly reputationally sensitive where it covers things like diversity statistics, equal pay, emissions and safety incidents. My GP clients say that data providers have to be able to build their own in-house LLMs [large language models] to allow for that high data security – they can’t outsource it.

“You would think that AI would democratise the market for ESG tools by giving hundreds of ESG data companies the ability to scale and provide services to GPs. My theory is the opposite: it is going to consolidate the market for service providers given that sample size of data required and the fact that you have to have the scale to be able to provide statistically relevant insights and extremely expensive data security.”

Humphrey also cautions GPs about the need to look at AI holistically when it comes to ESG strategy. “There are so many positives when it comes to streamlining investment into renewable energy and impact. I’m seeing some really interesting businesses at the intersection of ESG, sustainability and AI, for example, in areas like AI-driven environmental monitoring and climate analysis. The smart investors are really thinking about the risks and opportunities holistically.

“AI is another tool for innovation in private markets but, as with most new and disruptive technologies, there is both a light and a dark side. The positive is clearly AI’s ability to streamline processes and accelerate impact, but there is also the downside associated with AI’s exponential energy demand, job displacement and fair and ethical use issues. GPs need to be attuned to those multiple fronts from an ESG and sustainability
perspective.”