Bond Voyage

Read this article to understand:

• How AI is impacting the role of the research analyst
• What are its limitations?
• Is there still value in the human interaction?
   

“Will we still need research analysts in the future?” 

This is becoming an increasingly common question in our industry.

Given the advances in artificial intelligence (AI), it's a fair question. Financial analysts consistently rank among the occupations most exposed to AI technologies. According to a recent survey by the research company, Focaldata, the finance and insurance industry is ranked second – behind only tech – as the industry achieving the biggest AI productivity gains (see figure 1).¹
 

AI is already helping analysts gather information, summarise documents, draft reports and even build components of financial models. If machines can increasingly perform many of these tasks, what happens next?

While that question has sparked plenty of discussion, for us it is clear: others may view research as a cost to manage, but we see it as a source of investment advantage. High quality investment decisions depend on high quality research, and AI has not changed that belief.

Research is not just gathering information

History shows how the role of research analyst has evolved. There was a time when simply getting reliable company information was itself a competitive advantage. Technology transformed access to data, shifting the challenge from collection to organisation. As information ballooned, the job became increasingly focused on filtering, summarising and prioritising an ever-growing volume of news, data and commentary. AI is simply the latest chapter in that evolution.

Today, many tasks that were once time-consuming can be completed in seconds. Information gathering, spreadsheet population, transcript summarisation and first drafts of research notes can all be significantly accelerated. 

But those tasks were never the ultimate purpose of investment research.

Research is about interpretation and, crucially, generating insights that lead to better investment decisions. In fixed income markets, where pricing often reflects a consensus view of publicly available information, generating excess returns increasingly depends on interpreting information differently and identifying where market expectations may be too optimistic or pessimistic. Technology may alter how information is gathered and processed, but not the need for judgement in capital allocation.

Modelling has become easier

One area where AI is already changing behaviour is financial modelling. For many analysts, building a model from scratch was a rite of passage. But the value was never really in the finished spreadsheet. The real benefit came from the process itself. Manually entering numbers forced analysts to understand the business, highlighted inconsistencies, revealed key drivers and encouraged deeper investigation when something didn't look right. Importantly, it taught analysts which assumptions really mattered. 

But AI can now automate much of that process, which is not necessarily bad. The opportunity is to spend the regained time on what matters most: evaluating future scenarios, stress-testing assumptions and identifying what will ultimately drive returns.

As technology takes on more of the mechanical aspects of modelling, investment firms will need to think carefully about how those analytical instincts continue to be developed in the next generation of researchers.

Fluency is not the same as insight

The impact of AI on thinking is perhaps most evident in writing. Modern AI tools produce content that is often coherent, persuasive and highly polished. However, the challenge is that fluency is not the same thing as insight. 

Historically, writing has been one of the most effective ways for analysts to test the strength of their own reasoning, and arguments that appear convincing when discussed verbally, often reveal weaknesses when committed to paper. 

The act of writing helps sharpen priorities, reveal gaps in logic and show where evidence remains incomplete. An analyst who builds an argument gains a strong sense of which parts of the thesis are robust and which require more work. 

AI-generated content can obscure those distinctions as weak arguments are often presented with the same confidence and polish as strong ones. Thus, analysts need to be more deliberate in challenging and validating conclusions, rather than assuming that a well-written answer is a well-reasoned one.

Human interaction becomes more valuable

Perhaps the area least affected by AI is also the area becoming most important.

Research has always involved engaging directly with corporate management teams, sovereign policymakers, structured finance sponsors and a wide range of broader market participants. These conversations are not simply about collecting information. Strong analysts learn as much from how something is said, as from what is said, or even left unsaid.

They observe confidence, hesitation, consistency and credibility, identify where management teams appear uncomfortable, test assumptions and challenge narratives. Over time, the interactions build pattern recognition leading to a deeper understanding of strong management, warning signs that matter and strategies which may succeed or fail. That accumulated judgement is difficult to automate.

AI can help prepare for and summarise meetings. What it cannot do, is build experience through thousands of conversations across sectors, business cycles and market environments. The value of that experience rises as more routine research tasks become automated.

What remains scarce?

If AI makes it easier to gather, organise and summarise information, the investment advantage must come from elsewhere. The skills that remain difficult to replicate are the ability to ask the right questions, interpret conflicting evidence, challenge consensus views and ultimately decide which opportunities genuinely deserve capital. These capabilities are likely to become even more important as access to information becomes increasingly democratised.

So, when people ask whether we still need analysts in the future, they may be asking the wrong question. The more important question is: in a world where information is increasingly abundant, who is best placed to generate differentiated insight from that wealth of information?

The firms that carefully choose how to develop, test and apply AI will benefit the most from its possibilities. Those that mistake AI exposure for redundancy may find out otherwise, at the worst possible moment.

Should investors care?

The real significance of AI is not whether certain research tasks become faster, but how firms choose to use the capacity that technology creates. The most successful organisations will not simply use AI to produce more reports or cover more issuers, but free up time towards deeper analysis, broader opportunity sets and more rigorous testing of investment ideas. AI doesn’t reduce the value of research, but raises the standard of credible research.

As information becomes more abundant and the tools available to process it become more sophisticated, differentiation will come less from access to information, and more from the quality of human judgement applied to it.

When people ask whether analysts will still be needed in the future, they are often viewing the role through the lens of tasks rather than outcomes. Many individual tasks will undoubtedly change – some may disappear entirely. But the underlying objective remains the same as it always has been – to develop insights that lead to better investment decisions. 

Thus, the ability to generate differentiated insight becomes increasingly valuable. That is why the future of research is unlikely to be defined by AI replacing analysts. It is far more likely to be defined by the growing gap between those who use AI to enhance their judgement and those who rely on it to replace it.