BUSINESS & AI · OLIVER BRUCE
Agentic AI is here. And it changes how you work.
Oliver’s hands-on experience building AI agents for operations and inbound enquiries.
From tools to systems
Over a weekend, I started building with OpenRouter, Claude and OpenClaw. The shift was in how I thought about the work: less about individual tools, more about systems that can carry a workflow through.
Agentic AI describes systems that can perform defined tasks with less repeated manual input. That does not remove the need for oversight, especially when an agent can access business data or take action.
Two early experiments
My early experiments included an operations agent for diary management, email summaries and reporting, and an inbound lead agent to summarise enquiries, qualify opportunities, log information in our CRM and alert the sales team.
The opportunity is to reduce repetitive, low-value work so that people have more room for strategy, creativity and decision-making.
Structure before scale
Data access, permissions and security need careful consideration. A separate machine can help separate workloads, but it is not a substitute for clear access controls, human oversight and a reliable setup.
Start with a defined workflow, understand what the agent can access, and review its output before expanding its role. These are early experiments, not a guarantee of results for every business.