​​​​Beyond Numbers: How GenAI is Shaping the Future of Asset Management

Timothee Consigny, Chief Technology Officer at H2O Asset Management, was the firm’s first employee and has spent the last 15 years building the technology infrastructure of this global macro fund. Now, he’s embracing one of the most transformative technologies of our time: Generative AI (GenAI). Rather than rushing in, Tim has taken an employee-led approach to GenAI adoption. By empowering users and equipping them with the right tools, he’s demonstrated that allowing employees to take the lead can unlock innovations that drive the future of the business.

A Bottom-Up Approach to AI Adoption

As CTO, Tim empowered his colleagues to explore GenAI tools by using APIs to integrate existing large language models (LLMs), like ChatGPT and Gemini, with company messaging apps such as Slack. This bottom-up approach to AI adoption meant that instead of the management team imposing an AI strategy, the process was led by employees. As Tim puts it, “Innovation shouldn’t be driven by technology alone—it has to come from the end users and the business. With GenAI, we’re not just adopting a new tool; we’re exploring entirely new ways of working. Being a small boutique with 65 people gives us a real advantage—our experts sit right next to our tech team, allowing innovation to happen organically, together.”

By proactively providing access to GenAI, Tim mitigates the risks of Shadow IT, where employees adopt unapproved tools, potentially risking data breaches. As Tim explains, “The biggest risk for me is Shadow IT, where you don’t provide any solution for your employees, and they will just use their phone or go on Google and open the first Large Language Model (LLM) they can find, and you lose data privacy.” Tim’s approach strikes a balance between caution and trust. While ensuring data privacy through vetted tools, he resists the urge to monitor employees’ every move. “Innovation thrives on freedom, but it also needs a clear framework. Rather than imposing rigid controls, we established a high-level AI manifesto—simple, guiding principles like always keeping a human in the loop, not using AI for trading applications, and avoiding consumer-facing use. This approach allows employees to explore and experiment while ensuring AI adoption remains responsible and aligned with our core values.”

From Meeting Minutes to Eliminating Bias

By proactively providing employees with access to GenAI, H2O’s investment team uncovered a standout use case. Portfolio managers (PMs) at H2O Asset Management hold weekly meetings to discuss the market and their positions. Initially, GenAI was used to transcribe meeting minutes, but this quickly evolved into the creation of a valuable new dataset.

By feeding these transcripts back into the system, the team used GenAI to identify biases in their investment process. As Tim explains: “We asked the LLM, ‘Okay, do we have any bias as investors? Is someone consistently taking a contrarian stance? Are we stuck in repeating the same views without adapting to new information?’ This gave us a structured way to challenge our own thinking and bring a fresh perspective to our investment discussions”.

And one of the biggest advantages is GenAI’s impartiality. Tim adds, “If you ask a junior analyst to point out your biases, it’s not exactly easy for them to tell you that you’re making mistakes. But the AI is not scared of being fired or being unplugged (at least not yet!).”

The Future of GenAI in Asset Management

Tim believes we’re in the midst of a technological revolution. As he explains, “One of GenAI’s most remarkable capabilities is its capacity to handle abstract concepts and ideas.” 

In asset management, Tim sees GenAI complementing traditional quantitative models by scaling qualitative analysis in a way that was previously impossible. Take, for example, a scenario where you’re trying to assess the state of an economy. Sure, you can look at GDP, inflation, or unemployment rates and run your models, but you’re losing so much qualitative data that would offer you a much more holistic understanding of what’s occurring. “For decades, our industry has relied on quantitative models that provide scale and a semblance of certainty but often miss real-world nuances. By reducing companies to balance sheets and risk to price volatility, these models overlook crucial qualitative factors. GenAI broadens our perspective, allowing us to process concepts without forcing everything into numbers.”

When asked about AI’s potential in trading, Tim highlighted two critical challenges: alignment and hallucination. “When you tell the machine to do something, it will try to do that regardless of constraints,” he cautioned. In trading, this could result in AI blindly chasing profits through market manipulation or producing highly convincing but entirely fabricated market analyses. Tim also underscored the importance of keeping humans at the helm of trading: “I’m very sceptical of fully autonomous, agentic AI—especially in the context of trading. AI isn’t here to replace us; it should empower us to excel at what we do.”

Advice for fellow CTOs

Tim’s philosophy for adopting AI is clear: know your boundaries, involve humans, and prioritise accountability.

“The main challenge with AI isn’t the technology—it’s culture and governance. Everyone will have access to these tools, but only a few will have the right people interacting with them in a way that truly generates value. That’s why it’s critical to get everyone around the table early and agree on what you want to do with GenAI as an organisation—and, more importantly, what you don’t want to do with it. 

Before implementing AI, we set clear guidelines: always keep a human in the loop, never use AI for consumer-facing applications, and avoid any trading applications that could directly impact asset prices. AI is a powerful tool, but without the right framework, it’s easy to lose control of how it’s used.”

Making GenAI Work for You

Overall, Tim highlights that the key to successfully adopting GenAI is playing to your strengths. Every business has its unique advantages, and at H2O Asset Management, their strength is being a boutique firm. Owing to their size, Tim benefits from close collaboration with the investment teams. “I sit next to our CIO, who has 30 years of market experience. He helps me train our models, and the wider investment teams provide me with relevant data sources to explore the AI with, rather than just all the data possible.” This close-knit approach ensures that the AI is not only powerful but also aligned with the team’s deep expertise, making the technology work for them.

Looking for more insights?

Get exclusive insights from industry leaders, stay up-to-date with the latest news, and explore the cutting-edge tech shaping the sector by subscribing to our newsletter, Commodities Tech Insider.