AI agents are set to transform the job market, and financial analysis is the first target
AI agents are not just another productivity tool. They represent a structural shift in how companies operate, how professionals work, and how decisions are made in real time.
While traditional AI already does plenty of useful things day to day, like summarizing emails and organizing meetings, agents go further: they execute complex tasks autonomously, make decisions along the way, and function more like a coworker than an assistant. It might seem like a subtle distinction, but it changes everything when you start thinking about the real impact on work as we know it today.
That was exactly what Jai Das, co-founder and president of Sapphire Ventures, a venture capital firm with over 10 billion dollars in assets, talked about at the AI Agent Conference held in New York on a Monday. At the event, Das delivered a clear and direct take on where we are, what is already happening, and what is coming next, without excessive hype but also without underestimating what is at stake.
Spoiler: financial analysis is on the front lines of this transformation. 🎯
What AI agents do differently
The big shift that AI agents represent is not about the intelligence itself but about the ability to act. Generative AI tools, for example, respond when you ask. An agent does not wait to be called: it monitors, decides, and executes, all within a workflow that can involve dozens of interdependent steps. Think of it like this: you set a goal, and the agent maps out the path, solves the obstacles in the middle, and delivers the result. It is like having a junior teammate who knows what needs to be done without needing a manual for every situation.
According to Das, this operational autonomy is what puts agents in a completely different category. In practice, they can integrate systems, access data in real time, interpret contexts, and make micro decisions throughout the entire process. In the financial sector, this is particularly powerful because financial analysis depends on exactly that combination: speed, accuracy, and the ability to process massive volumes of data simultaneously, something no human analyst can do at the same scale and pace.
The point Das emphasized is that, at this early stage, agents still need a human in the loop to verify what they are doing. They amplify what a professional can do, functioning as an extension of human capabilities, but oversight remains essential. An analyst who used to spend days consolidating reports can now focus on the strategic interpretation of data because the agent already handled the operational part. This shift in the professional’s role is what makes this moment so relevant for thinking about the future of work.
Agents that write, review, and execute code
Das also mentioned that users are already building agents capable of writing code, reviewing that code, and executing it, all in a chained sequence. This capability shows how the technology is evolving from isolated tasks to complete workflows. It is no longer about answering a question or generating a snippet of text. It is about owning an entire process from start to finish, with internal checks along the way.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Das predicts that as these agents prove their reliability, the proportion of work delegated to them will grow. Trust is the key factor here. Nobody is going to hand a critical process to a system that makes frequent mistakes. But as results keep stacking up and best practices are defined in real time, the trend is for adoption to accelerate significantly.
Financial analysis in the crosshairs of intelligent automation
The word automation has become almost a cliche in tech discussions. But what Das presented at the conference goes way beyond what we usually associate with that concept. It is not just about replacing repetitive tasks with scripts or simple bots. Modern AI agents are capable of handling non-linear workflows, meaning processes where the next steps depend on the results of the previous ones and where surprises are the rule, not the exception.
In financial analysis specifically, Das pointed out that agents are already performing tasks like compiling data from multiple sources, cleaning and structuring that information, building financial models, and generating forecasts or scenario analyses. That kind of work would normally consume hours or even days of a human analyst’s time. With well-configured agents, the time drops dramatically and the margin of error shrinks.
This level of automation is already being implemented at major investment funds and fintechs around the world, and the results are starting to show up in response speed, error reduction, and the ability to scale operations without necessarily growing the team at the same rate.
Das was blunt when he said that companies are going to need to adapt. Not as a generic warning, but as a statement based on what he is already seeing in the market. Organizations that rely on large analyst teams for tasks that agents can execute efficiently are about to face competitive pressure that did not exist before.
The cost of keeping agents running
But it is not all sunshine and roses. One of the biggest challenges Das brought up during the conference was the issue of costs. Because AI agents often operate continuously in the background, running for hours or even days, compute bills can climb fast. Really fast.
Das was emphatic when he said that cost is going to be a major problem. This is a variable that many companies have not fully accounted for yet. Deploying an agent might seem cheap at first, but keeping it running at scale and around the clock requires robust infrastructure and financial planning that not everyone is ready to cover. This is a point that deserves special attention from anyone evaluating the adoption of this technology right now.
AI-native startups and the new business logic
One of the most thought-provoking parts of Das’s talk was his analysis of what he called a bifurcation among startups. On one side, there are AI-native startups, companies built from day one with agents and automation at the core of their business model, not as a layer added later. On the other side, older startups with larger teams of human engineers trying to make up for lost time.
AI-native companies operate on a completely different logic than traditional organizations. They can scale without growing headcount linearly because a large portion of operational work is handled by agents. This radically changes the cost structure, the speed of growth, and how value is created.
Das mentioned that these companies also adopt different pricing models. Instead of charging a flat monthly subscription for software access, many of them sell outcomes. The customer pays for the number of tasks the agent completes or for the actual usage of the platform. This outcome-based or usage-based pricing model reflects the shift in mindset: the value is not in accessing the tool but in what it concretely delivers.
According to Das, the key for a company to become truly AI-native is having a CEO focused on that transformation. Without committed leadership, agent adoption stays superficial and does not generate the results the technology makes possible.
Not everyone will survive
Das did not hold back on his view of the competitive landscape. He said outright that not every company is going to make it. According to him, that is simply how technology works. The ones that will stand out are those willing to take a radical approach, redesigning their entire organizations around AI agents.
This is not empty motivational talk. It is a market read from someone who manages a firm with billions in assets and closely follows hundreds of tech companies. When someone in that position says the transformation is going to separate those who survive from those who get left behind, it is worth paying attention. 🚀
What this means for people who work with data and finance
For professionals in financial analysis, data management, or any field that depends on processing large volumes of information to make decisions, Das’s message was clear: the question is not whether AI agents will change your job, but when and how each professional will position themselves in the face of that change. Those who understand how to work alongside agents, knowing when to trust them, when to question their outputs, and how to interpret the results they deliver, will have an enormous advantage.
Automation does not eliminate the need for human expertise, but it shifts where that expertise is most valuable. Mechanical and repetitive tasks tend to be absorbed by agents. What remains, and what will be worth more and more, is the ability to think strategically, to understand the context behind the numbers, to ask the right questions, and to translate agent insights into real decisions. Those who develop these skills will find a job market that is more interesting, not necessarily smaller, but fundamentally different from what we know today.
Customer service and programming are also on the radar
Although financial analysis was the big highlight, Das also pointed to other areas where agents are already proving exceptional: customer service and programming. In these fields, the agents’ ability to follow instructions, adapt to feedback, and improve over time is generating results that are turning heads. Customer service chatbots that actually solve complex problems, and coding tools that go beyond autocompleting lines, are concrete examples of what is already working.
According to Das, one of the most relevant observations right now is realizing just how good agents and models are getting. The evolution is not gradual: it is accelerating. And each improvement builds trust, which in turn increases adoption, creating a virtuous cycle that will make the technology increasingly present in everyday professional life.
We are at the beginning, and that is the most important thing
Das made a point of providing context: we are in the early days of AI agent development. That means the most powerful systems are still experimental. The basic capabilities already exist, but best practices are being built in real time as more companies test, fail, adjust, and share what they have learned.
This early phase is both challenging and full of opportunities. Challenging because there is no ready-made playbook to follow. Each organization needs to experiment, understand its own workflows, and figure out where agents generate the most value. But it is full of opportunities for that exact reason: those who start exploring now will accumulate knowledge and experience that will be extremely valuable when the technology matures.
The most accurate analogy might be the early years of the commercial internet. Back then, a lot of people did not understand the impact the web would have on business. The companies that experimented early, even making mistakes along the way, built competitive advantages that proved decisive in the decades that followed. With AI agents, the pattern looks the same.
AI-native startups are already playing by these new rules. Established companies are trying to adapt. And professionals across all fields, especially in financial analysis, are discovering that the ground beneath their feet is shifting faster than many imagined. The question is not whether this transformation will happen. It is already happening. ⚡
