AI Impact HubAI agents
A chatbot gives you something to action. An agent takes the action.
For anyone whose feed keeps shouting about agents without ever saying what one is.
Try it right here ↓Free, like everything in the Hub. No signup to read it.
People search for agents expecting magic or menace. Kyle's definition is calmer and more useful: you give an agent a goal, context and tools, and you review what it does — the same things you would give a new staff member. Which means you already know how to work with one. Like a manager.
The working difference from everything that came before: you give an agent the goal, not the steps. Kyle's example from his own week — he needed every PDF downloaded from a website, said exactly that and nothing more, then watched it try one technique, fail, try another, and keep going until the job was done. You set the outcome and the boundaries; it works out the route.
The shift that matters is not technical, it is managerial. In Kyle's words: we are moving from being the creator to being the editor — and his standing advice, as strange as it sounds, is to learn management skills, because we will be managing agents and agentic workflows. Knowing what good looks like, spotting what is wrong, asking for changes in a structured way: that is the skill set now.
The philosophy that keeps it safe is the one he repeats in every session: human in the loop. And there is a mechanical reason it matters where you stand in that loop — if you are only involved at the very end, there is exactly one check, and anything that went wrong earlier has compounded by the time it reaches you. Early on, build the check-ins in more often.
The place to start is smaller than the hype suggests. Take one task you would offload to a person. What would the best person for that job need to know? What tools would they need? What would their position description say? Configure exactly that. One agent, one job.
A chatbot is like talking to a very supportive friend: they guide you, give you ideas, a draft of what you can do — but you need to take that action. An agent actually has the ability to take that action on your behalf. It's like you've hired an assistant and given them access to your tools, to your environment. We're giving it hands.Kyle, opening his most recent AI agents webinar, August 2026.
- IntelligenceThe underlying model — the brain. You rent this part, and you can swap it as models improve.
- ContextWhat it knows about you and your organisation — the onboarding. An agent without context is an intern on their first day, sent off with no induction.
- ToolsWhat it can actually touch — email, calendar, files, the web. Kyle's image: the harness is a body around the brain, and tools are the limbs, moving only with permissions you grant.
- SkillsIts position descriptions — how your organisation does each task, written down once so the agent does it your way every time.
- ScheduleWhen it runs — every morning, every Friday, when something arrives. The schedule is what turns a chat into a colleague.
Kyle's five-part anatomy, the way he walks a room through it. Step through the parts, then take the design prompt at the end.
The model underneath — the brain. It already went to every university and read the whole internet. That part is solved, and it is the part you rent.
The intelligence is the least interesting part. Everyone has the same brains available.
Next: Projects.
A project takes your outputs from zero to 80%.
Read Projects →