In this episode I speak with Paul O’Sullivan, SVP of Solution Engineering for the UKI and Chief Technology Officer for the UK, from Salesforce. Paul traces 75 years of AI development, from early machine intelligence ideas to the breakthroughs that brought us to today’s agentic AI era. Paul explains the shift from traditional workflows to digital workers operating alongside people. AI-enabled reasoning, automation and data-driven decisions now open space for employees to focus on meaningful tasks rather than repetitive activity.

Paul outlines the elements senior managers must consider when forming an AI strategy. These include ethics, workforce readiness, data quality, resilience and long-term sustainability. He stresses that many AI pilots collapse because they lose focus on business value or fail to carry teams with them.

Data chaos is common in many organisations and progress is possible by concentrating on workflows rather than attempting a huge enterprise-wide data project. Paul explores the dark side of AI, including malicious code, deepfakes, identity theft and large-scale misinformation. Boards need clear oversight and consistent guardrails.

Paul discusses the build–buy–hybrid debate. The rapid cycle of AI development, model updates and new controls means few organisations can produce everything alone. Choice of partners becomes critical.

We end the conversation with Paul’s view of near-term AI advances: robotics with touch-sensing algorithms, multimodal models, self-adjusting systems and strong growth in sectors such as logistics, healthcare and financial services. In two years from now, will the landscape be almost unrecognisable?