I had the pleasure of guest lecturing at Sciences Po last Friday on how I’ve used political risk throughout my career — not only to analyse what politics means for investors, but also to build sustainable investment funds.
One student asked a great question:
“How have recent advancements in AI changed the outlook for jobs like yours in sustainable investing?”
As it happens, she asked this during a week when I had started experimenting with agentic AI for fund construction, so the question was already front of mind.
Given AI’s profound implications for many white-collar professions, it felt worth exploring further in this week’s edition of The Resilient Investor.
TLDR:
The Game Has Changed
First, a caveat: I am a user of technology, not an expert in it. I adopt tools when they make my work easier and more efficient, but I was not an early AI enthusiast — largely because I lacked the technical background to fully appreciate its potential.
If you had asked me this question a year ago, I probably would have described AI as little more than advanced Google — delivered with the enthusiasm and inaccuracy of an over-excited intern.
Six months ago, I would have said it was a powerful editing assistant that required close supervision due to misquotes and misattributions.
Two developments have since shifted my perspective.
First, the standard AI available through paid subscriptions (in our household, ChatGPT and Claude) has become materially more powerful and usable over the past year. Graphics that once took hours in PowerPoint now take minutes. Editing and sourcing are far more accurate — and improving. It can even offer useful perspectives on positioning and social media strategy.
Second, the emergence of agentic AI has expanded what AI can do exponentially — including in sustainable and thematic investing. Traditional ChatGPT-style interfaces operate on an ask-and-answer basis. Agentic AI, by contrast, can execute multi-step projects, use tools, and operate software. When I asked ChatGPT to summarise the difference, it likened itself to a calculator — while agentic AI is more like an associate who can work semi-independently.
For those interested in the mechanics, there is excellent material available online. The New York Times’ The Dailyrecently ran an episode titled “Can AI Already Do Your Job?” that is worth a listen.
But here, I’ll focus on what generative and agentic AI mean today for sustainable investment research and fund design.
A Note of Humility
Just as I underestimated how quickly generative AI would become useful, I am humble enough to admit I cannot foresee how far — or how fast — it will develop from here.
The speed at which these tools have penetrated markets and improved is striking. Ignoring them is not a viable strategy.
Those who understand how to work with AI — and think creatively about its application — will be better positioned to build innovative research and investment products. Staying alert and adaptable to change is key.
More Innovation, Streamlined Workflows
AI arguably creates opportunities to innovate and streamline at every stage of sustainable investment. The student’s question was in the context of fund creation and maintenance so let’s focus on that.
Building a sustainable investment fund typically requires:
- A clear thematic framework or taxonomy to guide decision-making.
- A wide range of data sources and management platforms to identify opportunities and risks, assess thematic alignment, and meet regulatory reporting requirements.
- Continuous top-down and bottom-up review to ensure consistency and integrity.
Agentic AI has the potential to drastically reduce workload across all three.
I remain convinced that human inspiration is essential to develop a fund’s theme and philosophy. But once that framework exists, agentic AI can meaningfully accelerate strategy construction and back testing.
For example, I am currently developing a potential fund framework that uses agentic AI to translate my thematic architecture into a portfolio — leveraging its coding, data processing and multitasking capabilities. This is work I could not realistically have undertaken alone even a year ago.
The process will undoubtedly require refinement. But it opens the door to experimentation and product innovation at a pace and scale previously unavailable to individuals and firms.
I remember spending many painful hours at university learning to code in econometric software called Stata. Then I joined the workforce — and no one was using Stata. Instead, it was EViews, or R, or Python, depending on the firm’s budget and security preferences. Each came with its own language, syntax, and quirks to master before you could do anything useful.
Today, that barrier is rapidly disappearing.
With AI, it is no longer necessary to know multiple coding languages simply to operate different systems. I have been genuinely struck by the speed and competence with which tools like ChatGPT and Claude can write code. It is no longer essential to sit deep in large datasets, manually wrangling software to identify, transform, and extract data points.
In this new environment, the scarce skill is not syntax — it is vision and creativity.
What matters is having a clear vision of the problem you are trying to solve, knowing how to instruct agentic AI effectively, responding intelligently to the questions it raises, and reviewing its output with enough judgment to catch errors.
That process is still rigorous — but it is far less time-intensive than the old model of doing everything manually.
This, in my view, is where AI’s greatest opportunity lies in sustainable investing: lowering data-heavy, time-intensive barriers to creating new, exciting investment products and enhancing existing products.
The innovation opportunity is clearest in passive and rules-based fund structures. However, active managers can also benefit — using agentic AI to streamline data surfacing, management, and reporting through multi-step automated workflows. Human oversight remains essential, but the operational burden is significantly reduced.
The result: greater innovation and more efficient operations.
The Workforce Implications
These shifts have real consequences.
Sustainable investing has already undergone significant cost-cutting over the past year as ESG appetite has cooled.
I believe AI and especially agentic AI is likely to mean fewer data gathering and processing roles in the sustainability space. Reporting is also likely to see substantial automation, impacting the need for roles in this space.
Junior hiring is the biggest concern. Firms must rethink how they build talent pipelines when traditional entry-level analyst tasks — chart production, model updating, data collation — are increasingly automated.
And yet, I am confident that AI will also create roles we cannot yet fully define.
Creative, thoughtful professionals will be needed to design systems, interrogate outputs, refine frameworks, and build products that responsibly integrate AI into investment processes.
Those who embrace change — and focus on delivering better outcomes for clients — will find ways to harness these tools in ways we can scarcely imagine today.
If you enjoyed this piece, subscribe using the button below. I also welcome feedback and questions at [email protected].