In late November, a start-up called “/dev/agents” captured headlines with its $56 million seed funding round at a $500 valuation. Not bad for a company without a commercial product. It’s easy to see why the buzz was electric: the founding team boasts alumni from Google, Stripe, and Meta Platforms.
The company is building an operating system for AI agents. These agents are poised to redefine how we interact with software. The idea is that they run in the background, acting autonomously based on user-provided guidelines and with minimal human oversight.
AI agents have become a hot pursuit of tech giants. At Microsoft’s Ignite conference, CEO Satya Nadella sang their praises, while Salesforce rolled out Agentforce, its take on this transformative technology. Not to be outdone, ServiceNow, Workday, and Oracle have joined the race with their own AI agent offerings.
“The rapid adoption of AI agents stems from their transformative impact on productivity — automating repetitive tasks, accelerating incident resolution, and freeing up valuable time for innovation,” says Karthik SJ, who is the general manager of AI at LogicMonitor. “This shift is especially visible in how these AI agents help teams interact with their systems through natural language, making complex operations more accessible.”
It’s still early days for AI agents. Yet the potential is clear for many industries, including financial advice.
What are AI agents, really? That depends on whom you ask. There isn’t a universally agreed-upon definition. But an AI agent isn’t an AI model like OpenAI’s GPT-4 or Anthropic’s Claude 3.5 Sonnet.
“An AI agent is a system that packages AI models,” says Madison Faulkner, a principal at NEA, which is a leading venture-capital firm. “It’s done in a way similar to how a person operates. This is why there’s such a big unlock with use cases.”
An AI agent stands out because it brings a unique set of capabilities to the table. Think of ChatGPT as an example. At its core, ChatGPT qualifies as an AI agent because it interacts with the real world by processing user prompts and generating responses in natural, conversational language.
But there’s more to it than just answering questions. ChatGPT also uses tools. Ask it to create an image, and it taps into DALL–E. Need some code or help navigating the web? It can handle those too. What’s more, ChatGPT operates with a level of autonomy. Once you provide a prompt, it takes over and completes the task.
So, why not stick with ChatGPT for everything? The next generation of AI agents are designed for specific industries and domains.
“Sophisticated AI agents use specialized large-language models,” says Abhi Maheshwari, CEO of Aisera, which is a developer of AI agent software. “They are based on a company’s proprietary data and industry information. This makes the responses much more personalized, accurate and useful.”
AI agents will have access to more tools, especially those that are common for businesses. These include email systems, collaboration tools like Slack, ERPs and CRMs.
But it doesn’t stop there. These agents will also adapt and improve over time. Feedback from users won’t just inform their behavior — it will actively shape it. In this way, AI agents will evolve alongside the businesses they serve.
Financial applications. Moody’s may have over a century of history, but the company is far from stuck in the past. The financial giant has been making bold moves to develop and deploy AI agents.
“We’re seeing fundamental improvements in both LLMs and agent frameworks that allow us to build enterprise-grade applications,” says Sergio Gago Huerta, who is Moody’s managing director of AI and machine learning.
The potential for AI agents in delivering financial advice is undeniably impressive. In an industry where vast data sets and complex workflows are the norm, these tools offer a powerful means to streamline processes and enhance efficiency.
“One of the most compelling applications of AI agents is the ability to help financial advisors better inform their clients on financial decisions,” says Babak Hodjat, who is the CTO of AI at Cognizant. “A set of agents can work together to analyze a client’s financial history, spending patterns, and investment portfolio to generate tailored recommendations. For instance, one agent might focus on creating a personalized budget plan, while another identifies investment opportunities that align with the client’s risk appetite and long-term goals.”
Yet financial advisors should not fear that their role will be disrupted. “AI Agents and other AI related tools are likely to accelerate the evolution in the advisor’s value from technical expert to the financial ‘guide’ for their clients helping them to understand their true needs and find solutions to meet their needs,” says Craig Martin, who is the global head of Wealth & Lending Intelligence at J.D. Power. “The human advisor’s primary advantage over AI is the personal relationships and understanding of their client’s wants and needs.”
Getting Started. Financial advisors looking to take the first step with AI agents can start with established platforms like Salesforce and ServiceNow. These platforms come with robust data access, enterprise-grade security and prebuilt workflows.
Looking ahead, the landscape is likely to evolve rapidly. In just a few years, we can expect to see a growing array of AI solutions designed specifically for the unique needs of financial advisors.
Tom Taulli ( @ttaulli ) is a freelance writer, author, and former broker. He is also the author of the book, Artificial Intelligence Basics: A Non-Technical Introduction .