NextCorps recently hosted a workshop with Chris Rimlinger, owner of New Market Ventures, LLC, who offered a deep dive into one of today’s most in-demand skills: how to craft clear, effective prompts that actually deliver useful results from AI.
Rimlinger’s workshop was designed to help professionals move beyond trial and error with large language models (LLMs) like ChatGPT, Gemini, Claude, and Perplexity. Attendees learned practical frameworks and advanced prompting techniques that turn vague requests into precise directions to ensure better outputs—whether they’re writing content, analyzing data, or brainstorming ideas.
When AI works, and when it doesn’t
Throughout the session, Rimlinger reminded participants that AI can be powerful, but it isn’t magic. It’s best used as an assistant, not a replacement for true expertise. Tools like ChatGPT shine at brainstorming, summarizing, translating, or providing creative direction. They can also assist with code review, debugging, and role-play scenarios, such as customer conversations or business negotiations.
However, Rimlinger warned against relying on AI for legal, medical, or financial advice, or anything that demands precision or privacy.
“AI can be confidently wrong,” he noted, explaining that while these models often sound certain, they don’t actually know anything. Users should always verify factual or critical content and never input sensitive information into free versions of these tools.
The five golden rules of prompting
Rimlinger broke down five simple but powerful rules that separate good AI interactions from great ones.
- Be specific: Vague questions lead to vague answers. Define exactly what you want and how you want it delivered.
- Give context: Include information about your audience and goal. This helps AI tailor its tone and depth of response.
- Set a role: Assign expertise, such as, “You are a startup advisor,” or “You are a technical writer for a healthcare audience.”
- Iterate and refine: Treat prompting as a dialogue. Build on each response to reach better outcomes.
- Break down complex tasks: Instead of combining multiple steps in one command, guide AI through them one by one.
Rimlinger emphasized that prompts are more than simple requests. They’re the instructions that teach the AI how to think about a problem. The clearer your guidance, the better your results.
Rimlinger also shared what he called a golden follow-up prompt: “How could I have made that last prompt better?”
This question invites the model to act like a coach, offering suggestions to strengthen your prompting technique. Over time, this iterative method helps users internalize what makes a clear, purposeful request.
Turning prompts into a process
Using real-world examples, Rimlinger demonstrated how to transform everyday requests into structured prompts. For instance, a simple instruction like, “Write about marketing trends” might produce a generic answer. But reframing it as, “Write a 300-word article about digital marketing trends for small business owners, focusing on social media strategy” will result in a more relevant, focused response.
Once attendees mastered the fundamentals, Rimlinger introduced advanced techniques that take prompting to the next level:
- Few-shot prompting involves showing the AI two or three examples to teach it a desired tone or structure.
- Chain-of-thought prompting asks it to think step-by-step to reach a conclusion, improving accuracy for analytical tasks.
- Role stacking simulates multiple perspectives—like having a marketer, designer, and customer “discuss” a campaign—to reveal insights from different angles.
He also explored prompt chaining, a method of building multi-step workflows. For example, when developing a marketing solution, you might start by identifying challenges, then ask for potential fixes, format the results in a table, and finish by summarizing them in executive-ready language. By layering prompts intentionally, you turn AI into a structured problem-solving partner.
The human element
Despite AI’s growing sophistication, Rimlinger reiterated that human judgment remains essential.
“AI doesn’t replace critical thinking,” he said. “It’s an accelerator, but only if you know where you’re going.”
Rimlinger encouraged users to bring curiosity, skepticism, and creativity to every interaction, verifying key facts and refining tone to align with brand or audience needs.
The message was clear: AI works best when you work with it, not against it. By shifting from vague commands to guided collaboration, anyone can unlock more consistent, insightful, and reliable results.
Want more expert insights?
Subscribe to the NextCorps newsletter to get notified about upcoming lunch & learns and other events.