Permutable Founder and CEO says the next phase of agentic AI in finance will be defined by supervised workflows, source traceability and decision intelligence – not unchecked autonomy
London, UK – 17 June 2026 – Wilson Chan, Founder and CEO of Permutable, has called for a more disciplined approach to agentic AI in institutional finance, arguing that its real value will come from supervised, auditable workflows rather than fully autonomous systems making unchecked decisions.
As hedge funds, asset managers, banks and fintech platforms explore how AI agents could support investment research, trading intelligence and risk monitoring, Chan warns that financial institutions must focus less on the promise of autonomy and more on whether these systems can be trusted in real decision environments.
“Agentic AI has enormous potential in financial markets, but the industry needs to move beyond the theatre of autonomy,” said Wilson Chan, Founder and CEO of Permutable. “The most useful applications will not be systems that replace human judgement. They will be controlled workflows that help teams process information more effectively, preserve context and make better decisions faster.”
Agentic AI refers to systems that can pursue a defined objective across multiple steps, using data, reasoning and workflow orchestration to complete a task. In institutional markets, this could support areas such as research monitoring, market intelligence, risk review and decision support.
Chan says these capabilities are especially relevant to macro and commodities markets, where price movements are often shaped by fragmented, fast-moving information across regions, languages and asset classes.
“Macro and commodities are strong use cases because the relevant signals are rarely found in one place,” Chan said. “A meaningful development may emerge through policy language, local market reporting, supply-chain disruption or geopolitical events. The challenge is understanding whether the information matters, which markets may be exposed and whether it warrants further review.”
According to Chan, institutional investors should ask several practical questions before allowing AI agents into live workflows:
- Can outputs be traced back to credible source material?
- Does the system separate evidence from inference? How is hallucination risk monitored?
- What happens when market conditions change?
- Can the workflow be reviewed after the fact?
- Where does human judgement enter the process?
- What controls exist around the data being used?
“Institutional investors need to treat these as governance questions, not technical afterthoughts,” Chan said. “They need confidence that an AI-generated output is grounded in evidence, can be challenged and can be explained to a portfolio manager, risk team, client or regulator.”
Chan identifies model drift, hallucination and weak audit trails as key risks. Model drift can occur when systems become less reliable as conditions change. Hallucination can produce unsupported outputs. Weak auditability makes it difficult to explain why an AI system produced a signal, recommendation or escalation.
“For institutional markets, a black-box AI workflow is not good enough,” Chan said. “In markets, speed matters, but speed without provenance is not intelligence. AI systems need to show enough of their working to support review, challenge and accountability.”
Permutable’s position is that agentic AI in financial markets is ultimately a data infrastructure challenge. The company builds market intelligence technology that transforms global news, macro developments, geopolitical events and sentiment signals into structured, source-linked intelligence for institutional teams.
“An agent is only as useful as the data environment it operates within,” Chan said. “If the data foundation is weak, the agent will simply automate uncertainty at scale. Trusted data infrastructure is what turns agentic AI from a productivity experiment into a decision-intelligence capability.”
Wilson Chan is available for interview, Q&A or expert comment on agentic AI in institutional finance, investment research, AI governance and decision intelligence.
See also:
Permutable Named Hedgeweek Technology Provider of the Year: Innovation 2026 | Disruption Banking
