This blog post was inspired by a conversation we had with a client that is currently going through a company-wide AI transformation and exploring ways to accelerate

GenAI is transforming the finance industry, and firms face a stark choice: become AI-first or get left behind. The pace of change is unforgiving. Early adopters are already embedding AI into every decision, research processes, and client interactions. Firms who hesitate risk waking up in a market where their competitors move faster, know more, and deliver more. AI is no longer a side project, it’s a core part of their business. Many asset owners are already asking how the fund managers they’re invested in are leveraging AI. 

Too many institutions are taking a “toe in the water” approach: buying an off-the-shelf tool for a handful of “innovation champions,” running a few pilots, and waiting to see results before scaling. It feels safe and prudent. But in reality, it’s a recipe for irrelevance.

Why partial and selective adoption fails

Buying point solutions for a few early adopters doesn’t change the way the firm works. It doesn’t make AI part of the daily workflow of every analyst or portfolio manager. It doesn’t lead to the necessary cultural shift that is required in AI adoption. Instead, it creates isolated wins with no compounding effect and no realistic edge.

Worse, it risks breeding skepticism internally. When AI adoption is limited to a few showcases, the majority of employees are left wondering whether the technology really matters to them, or at all. 

Point solutions are also limited in value. Something we often hear is

“We just want to be able to access public filings, we won’t be integrating our own data into Quantly”.

A point solution can definitely add value to the analyst process, but it only touches the surface of what’s possible. Integrating internal sources of data and building a complete data layer into a firm’s AI infrastructure should be a top priority. 

However, what we see is that by the time leadership decides to expand from point solutions such as the general AI chatbots, the early movers in the market have already built a reliable and scalable AI infrastructure, started unifying their data in a single layer, trained their teams, and integrated AI into the fabric of their tech stack and decision-making.

The most successful AI systems are ones that work as a single source of truth for the firm over time. This does not have to happen all at once, as bringing all the data in can be a resource intensive process, but it should be on the roadmap of every investment firm.

AI transformation must come from within

An AI-first strategy is not about buying the “right” tool, it’s about building the right foundation that allows firms to build new use cases and scale in a fast and reliable way. That starts with three imperatives:

1. Universal access for analysts – Every desk, every team member should have the tools to augment research, automate workflows, and generate insights in real time. Currently firms may have a few early adopters to which usage is limited. This will change over time and the success of adoption is mainly driven by top management.

   Management is instrumental in mandating that teams adopt AI and use it. Doing this also opens the door to analysts getting creative and suggesting new use cases and ideas. It accelerates the AI wheel. More on this in our next blog post.
   
   

2. Data readiness – GenAI is only as good as the data it can access. Firms need to connect internal and external sources securely. This is not as simple as connecting to an MCP. Often firms need to invest in getting their internal data LLM-ready in order to extract maximum value from it. Data labelling and knowledge graphs are areas of experimentation that show great promise.
   
   

3. Scalable infrastructure – The ability to build, deploy, and iterate on AI agents quickly and securely is non-negotiable. Without a flexible, compliant, and observable AI platform, scaling is impossible. Ideally, give end-users ability to build hyper-customised agents fit for their work. They are the domain experts.

AI as a Strategy Differentiator

Truly adopting GenAI and building it into a lasting technology edge can improve analyst and PM productivity, expand their observable universe of companies and bring in unique insights at scale. But it also changes how the market sees you. 

For asset managers and banks, a credible, AI-first strategy signals to asset owners that you’re operating at the forefront of innovation, with the speed and analytical power to generate better insights and returns. Institutional investors increasingly want to place capital with firms that have a clear plan for leveraging AI to stay ahead. 

In a crowded space, showing that you’ve embedded AI into the heart of your investment and operational processes can be the differentiator that wins mandates and deepens client trust.

“Wait and see” is really “wait and lose”

Laggards will have a difficult time catching up. Investment firms are in a race where the winners will be the ones that can adopt AI fastest, integrate it deepest, and continuously evolve their capabilities. They will be the ones that have committed to an AI-first approach and invested in building out a reliable foundation for AI. The key to this is how fast a firm can adopt the upcoming technological improvements and bring them into production. A timescale of months is already too long. 

Competitors who go all in now will develop a cultural AI fluency, a proprietary library of internal AI agents, and a pace of execution that late adopters will struggle to match.

Financial institutions are faced with a choice

Either commit to making AI part of your firm’s DNA or watch others define the future of your market. The real winners have to commit to an all-encompassing transformation, whether that’s working with partners to get there or building it internally. 

It's also important to keep in mind that AI is not magic. Large transformations take time and although bringing in ChatGPT for everyone is a first step to that, it can often deceivingly feel like it's enough and your firm is adopting AI successfully. It's not.