As AI becomes more capable, how much freedom should we give it, and who is responsible when something goes wrong?
In episode two of AI Matters, host Kath Lord-Green is joined by AI expert Andrew Bel-Dean to explore trust, agentic AI and why businesses need to understand the data and guardrails behind the technology they use.
The conversation looks at the shift towards agentic AI, where AI can go beyond answering questions to creating code, applications and tools designed around specific business needs.
Andrew explains how this could enable organisations to move away from generic software towards applications built around their own data, processes and challenges. From identifying potential cash flow pressures to analysing business performance and competitor activity, the opportunity is to make AI more relevant to the decisions organisations make every day.
The episode also explores why human oversight remains critical. AI makes decisions based on the data and instructions it is given, which means businesses need to set clear boundaries around what their systems can and cannot do.
Andrew said: “AI is built on data. It will make decisions based on that data, but it’s down to the human to make sure the guardrails are in place, not the AI.”
Key takeaways from episode two:
- Agentic AI is opening the door to applications built around individual businesses and their needs
- Organisations can use their own information to make AI outputs more relevant and useful
- Bespoke AI tools could help businesses analyse performance, anticipate challenges and make better-informed decisions
- Human imagination, judgement and oversight remain essential
- Clear guardrails are critical as AI systems become increasingly capable and autonomous
Listen to episode two of AI Matters to hear the full conversation.
