Fintech has emerged as a leading sector at the forefront of AI adoption and automation. But recently, there has been some confusion about what agentic AI is, and how AI agents can be effectively put to use on both routine and complex tasks.
Last month, we looked at how
Citi GPS is using cutting-edge agentic AI, which it has dubbed the ‘do it for me’ economy.
Unlike GenAI which relies on user prompts, agentic refers to systems of models that have the ability to make decisions, take actions, and pursue goals, imitating human agency. This is all done independently.
5 fintech jobs hiring right now
- Sales Director - FinTech, Techwave, Greater London
- Partnerships Assistant - Fintech/Payment, Guardian
Professional Corporate Services, London Area
- Commercial Lead (FinTech), LoanTube, Hammersmith
- Solution Architect, FintechO, London Area
- Contech Sales, IB Global Services, England
In fintech, agentic AI’s applications range widely. Think of autonomous fraud protection, or streamlining compliance processes such as know your customer (KYC), and anti-money laundering (AML). Agentic AI can significantly reduce both manual effort and error
rates.
Once you understand agentic AI’s uses, AI agents become all the more clearer.
Common uses of AI agents
Helpfully,
HubSpot CTO Dharmesh Shah recently shared insights that help demystify AI agents, presenting them not as a binary concept but as existing on a spectrum of capability and autonomy.
Fintech professionals may find this framework useful for understanding how agentic AI is poised to reshape daily workflows.
Shah classifies agents into three different types.
Firstly, there are conversational agents. Although they may look much like the poorly-functioning chatbot of yore, if you engage with a conversational agent, the differences are suddenly stark.
In fintech, these agents can handle complex customer queries. And if these agents use
conversational search, they’re sophisticated enough to understand context, remember previous conversations, and provide tailored and personalised responses.
By using NLP (natural language processing), conversational agents will also understand the intent behind the question.
If you’ve been job hunting recently, you may even have happened upon, and used, an AI agent already.
Take Robin for example. This AI-powered agent sits on leading companies’ careers pages, and gives prospective candidates information on roles, provides CV to job skills matching, and more, in a conversational way. It’s
also available 24/7 and speaks multiple languages.
Agentic AI-savvy HR professionals already know what a timesaver it is to have an agent like this take over early-stage queries and questions.
Next, Shah identifies workflow agents as the second type of AI agent, and these can orchestrate multi-step processes, from KYC procedures to fraud detection, and regulatory reporting.
As Shah notes, these agents can be triggered manually or automatically in response to specific events, such as new customer onboarding or suspicious transaction alerts.
Thirdly, there are hybrid agents. These combine conversational interfaces with traditional UI elements, offering a flexible approach that should prove particularly valuable in fintech environments where both human oversight and automated processing are critical.
Hybrid agents can manage complex processes while pausing for necessary human approvals, an essential point of difference for maintaining control in heavily-regulated industries.
Impact of AI agents
For fintech workers, the integration of AI agents provides some pretty immediate change. Early adopters should see enhanced productivity fairly quickly.
And though a couple of years ago, headlines screamed about AI replacing jobs, it now appears that Jensen Huang, CEO of Nvidia, may be closer to the truth. He is repeatedly quoted as saying: “The person who uses AI will take your job.”
AI agents are emerging as powerful collaborators. By handling routine, and let’s face it, dull aspects of your workload, you are freed up to focus on more strategic decision-making, complex problem-solving, and empathetic people management.
For agentic AI-skilled workers, roles may become heavily focused around orchestrating and supervising AI agents. Again, Huang has the timely quote: “The IT department of every company is going to be the HR department of AI agents in the future.”
This means that IT professionals will supervise “digital workers” i.e. onboard AI agents and ensure they are properly trained and managed, similar to how HR departments manage employees. Switch out “IT professionals” for “fintech professionals” and the
same logic applies.
Finance workers will need to develop skills in prompt engineering, agent configuration, and understanding the boundaries of AI capabilities.
Additionally, Shah hopes that in the future “agents will be able to discover each other and collaborate to accomplish higher-order goals”.
For now, look for easy integrations into your existing tools, and go from there. It may be a case of rebuilding or remapping your workflows completely in the end, but it helps to play around with different agents to get a feel for their benefits first.
In fintech, finding the right balance between automation and human oversight will be key. But as Shah emphasises, the ultimate goal is for agents to be "useful to humans and help us work better."
AI agents aren't just another tool to learn. They represent a fundamental shift in how work gets done, and these are set to shape the future of work, whether you’re ready or not.
Ready to find a new role? Visit the Finextra Job Board where you’ll find new roles updated daily.