The businesses getting the most from AI, according to Dr. Connor Robertson, founder of Elixir Consulting Group and host of The Prospecting Show, are not using more tools than their competitors. They are designing better workflows. Robertson breaks a workflow down into four components: inputs, triggers, logic, and outputs, which he says appear in every effective AI workflow, regardless of the process being automated.
Component 1:Inputs
Every workflow begins with data entering the system, in Robertson’s framework: an email, a form submission, a calendar event, a payment confirmation, or a file upload. He argues the quality and consistency of inputs determine the quality of everything downstream, and that inputs need to be standardized, same format, same required fields, same structure every time. A workflow receiving inconsistent inputs, he says, produces inconsistent outputs regardless of how well the rest of it is designed.
Component 2: Triggers
A trigger is the specific event that initiates the workflow. Not every input should trigger automation immediately, in Robertson’s view; some need to be batched, some need a human approval checkpoint before the next step executes. Defining the trigger correctly, he argues, prevents under-automation, where a workflow never starts because the trigger is too narrow, and over-automation, where it fires in situations it should not.
Component 3: Logic
Logic is the conditional layer between the trigger and the output: if this condition is true, execute this path; if this field contains this value, route to this sequence; if this threshold is crossed, escalate to this person. Robertson describes logic as the point at which judgment is encoded into the system; the more completely a business can document its decision-making process under explicit conditions, the more powerfully the logic layer can operate.
Component 4: Outputs
Outputs are what the workflow produces once all preceding components have executed: an email sent, a Slack notification fired, a CRM record updated, a document generated and saved, a task created and assigned. Robertson argues outputs should be actionable and delivered automatically to where action can be taken on them; a workflow output that requires a human to go retrieve it before acting, in his framing, is not a complete workflow.
A Real Example: Form Fill to Personalized Outreach
Robertson walks through a sample workflow: a prospect fills out a contact form (input); the submission fires a webhook to the automation platform (trigger); if annual revenue is above a defined threshold, the lead routes to a high-priority sequence, if below, to a standard sequence, and if the field is blank, it gets flagged for human review (logic); a personalized first-touch email is generated byAI using the form data as context and queued for a 60-second review before sending (output). He describes that as a complete AI workflow built in an afternoon.
Why Most Broken Workflows Fail at the Trigger, Not the Logic
Robertson observes that when a workflow he’s reviewing isn’t working, the fault is more often in the trigger than in the logic layer that gets most of the design attention. A logic tree can be flawless and still produce bad outcomes if it fires too early, too late, or on the wrong event entirely. He treats trigger design as the component most likely to be underbuilt relative to its importance, precisely because it’s the least visually interesting part of a workflow diagram; there’s no elaborate branching to admire, just a single decision about exactly when the process should start.
About Dr. Connor Robertson
Dr. Connor Robertson is an entrepreneur, author, and strategic advisor based in Pittsburgh. He is the founder of Elixir Consulting Group, host of The Prospecting Show, publisher of The Pittsburgh Wire, and founder of The Grant Finder. He is also a six-time published author, with titles including Built to Run, available at
drconnorrobertsonbooks.com. More on his work is available at drconnorrobertson.com.
