Audacity in the age of the machine.

Yes. Jacqueline V. Twillie has no computer science degree. She worked in marketing while the Yellow Pages moved from print books to digital directories, worked in global disease eradication at the CDC, ran her own negotiation and leadership firm for a decade, and served as co-CEO of an international NGO in Australia. When she started building no-code AI agents, her first instinct was self-doubt about teaching it without a technical background.

What she found: building an AI agent that works is about clear communication, systems thinking, leadership, and knowing how to delegate. Every transition she had already survived gave her the toolkit for this one. She has since taught more than 400 people with no code background to build their first AI agent.

Both, on purpose. The most common concern in Jacqueline's workshops is "I don't want AI training on my data." She answers with two sides. When women and Black and brown communities were left out of research, from car crash test dummies to medical trials, the systems built without their data failed them. If whole communities opt out of AI, the models of the future will not reflect their lives. At the same time, blind participation is not the answer. Be intentional about what you share.

Her practical step: create a dedicated email address for your AI tool accounts, so the inbox holding your banking and medical mail is never the one you grant an AI agent access to. A separate email organizes your access; it does not turn off model training or protect sensitive files you upload yourself. Check the privacy settings in each workspace, and never upload confidential records you are not prepared to expose.

When you move to a new city, you expect to burn a tank of gas getting lost. You miss an exit and take the next turn without deciding you are bad at driving. Jacqueline applies the same rule to learning AI: people hesitate to spend twenty dollars on a month of credits, or quit the first time a workflow flops. Treat it as a deliberate investment instead of an expense.

Three moves: set a fixed learning budget of time and money each month so you are not using tools without limits, treat a failed workflow as a missed exit rather than proof you are not technical, and review what worked before you set next month's budget.

Because that is your job security. Companies that expected AI to replace people outright are learning an expensive lesson between rising token costs and workflow errors; many of the layoffs blamed on AI were corrections for pandemic overhiring. AI agents do repetitive, tedious steps well: pulling standard data, pre-filling forms. Humans own the context, the quality check, the ethical judgment, and the final sign-off.

Jacqueline teaches every team to keep a human in the loop. Delegating the execution of a task never removes your responsibility for the decision. Let the AI agent fill the tedious forms; you bring the audit and the final call.

You do not need to read fifty AI headlines a day. Jacqueline picks one trusted source and returns to it every week. Coming from a non-technical background, she does not expect a new technical concept to click on the first pass; sometimes she reads the same article or watches the same tutorial ten times until it connects to skills she already has. She also blocks thirty minutes to read real research instead of repeating viral talking points.

The other half is stepping away. The weekend before the episode she set up her easel outside, mixed paint, and painted for hours. Constant tool use is not progress. Fluency comes from focused practice and from leaving room to be human.

Because an AI agent is a system-connected workflow that runs on instructions, and most people set one up the way they would never treat a new hire: no context, no access, one shouted sentence. Jacqueline's test is to onboard an AI agent the way you would onboard a competent new person: the job and its goal, the role and context, system access, step-by-step instructions with the why, guardrails, and reporting and review.

The full six-part onboarding sheet is on its own page, along with the five-minute exercise for writing your first AI agent job description. The steps up from basic chatting to an AI agent team are on the AI Staircase page.

The onboarding sheet · The AI Staircase · The Friday audit

Rule one: you are not allowed to say "I'm sorry" before asking a question. Too many tech spaces treat people as less than if they do not already know the jargon, and that is how an access gap widens. Rule two: you must teach one other person something you learned. The sessions are not recorded. Jacqueline spent years on the gender wage gap, and the data taught her that when an access gap balloons early it can take centuries to close.

The free sessions started as a casual invite to a few friends and turned into more than 200 women raising their hands. Rule two comes from a principle Aboriginal and Torres Strait Islander communities shared with her when she worked in Australia: you have to give it away to keep it. Her 48-hour challenge after every session is to explain one takeaway to a friend, colleague, or family member over coffee or text.