ISITC TALKTIME: AI Agents in Financial Services
AI Agents in Financial Services: Opportunities, Risks and Human Oversight
When AI moves from answering questions to taking action, how do you keep control?
Join Gary Wright, Director of ISITC Europe, in conversation with Bob Morel, founder of AIRTA Systems for a plain-speaking exploration of AI agents in financial services. This episode examines what separates agents from large language models, why their ability to act raises governance questions, and how human oversight can support innovation without surrendering accountability.
Beyond the chatbot: why AI agents matter
A useful way to understand the discussion: the language model provides the thinking, while the agent supplies the ability to do work. The speakers explore how access to organisational tools and systems changes AI from a source of answers into a participant in a wider workflow.
But greater capability brings a bigger question: what should an agent be allowed to do on your behalf—and when should a person step in?
What you’ll hear in this episode
- AI agents versus LLMs: the difference between consulting a model and giving an agent an objective, tools and the ability to act.
- AI governance and compliance risk: why an agent making commitments or taking actions needs clear boundaries.
- Explainability and observability: the importance of understanding an agent’s actions and tracing the information behind its decisions.
- Human in the loop versus human on the loop: approving decisions before action, compared with monitoring activity and intervening when needed.
- Trading limits and escalation: how existing financial-services controls offer a starting point for thinking about agent permissions and approval thresholds.
- Market-wide risks: the speakers’ concerns about feedback loops, AI-generated data and the possible cumulative effects of automated trading.
Who should listen?
This episode is for financial-services leaders, risk and compliance professionals, technology teams and anyone asking how to put agentic AI to work responsibly. Expect accessible analogies and a candid debate, not a technical implementation guide or a promise that every risk has been solved.
Innovation without losing control
The conversation centres on a practical tension: approving every action can slow a process down, while unchecked autonomy can weaken control. Hear the speakers explore a combination of defined permissions, human approvals and ongoing monitoring—and ask how those safeguards could fit alongside established risk-management mechanisms.
The episode also looks beyond individual firms to ask what might happen when AI-driven activity interacts across markets. These are exploratory concerns and opinions raised in the conversation, not evidence that AI is manipulating markets.
Listen and join the conversation
How much autonomy is enough? Where should an agent’s authority end? And who remains accountable when it acts?
Listen to Part One and explore the opportunities and risks of AI agents in financial services. Share the episode with colleagues working on AI adoption, governance or trading controls. The discussion continues in Part Two.
Viewing time: 22 mins
Part 2 – Continues the discussion and will be published next week.
Bob Morel is the CEO of AIRTA Systems, an AI Risk Management platform designed to facilitate market access for AI-driven applications, and Co-Founder of Genbounty.
With a robust background in thought leadership, Bob previously served as the Head of Application Security at Centrica and the Application Security Lead for a cryptocurrency exchange. He is an active contributor to the cybersecurity community as an author for Infosec, creating learning paths on topics such as Breaking LLMs and AI-driven Applications, and the ChatGPT for offensive security. His expertise is supported by a B.Sc. in Computer Science, and industry certifications including the (ISC)² CISSP, CompTIA Security+, and SecAI+.
Leave a Reply
You must be logged in to post a comment.