These are my notes from a hands-on workshop on using ChatGPT and other AI tools for content creation and business - the practical patterns I actually use, not the hype.
A quick note on dates: this guide started life in July 2023. The tools have moved on enormously since, but the patterns - how to prompt, where AI helps, where to keep a human in the loop - have held up remarkably well.
These tools are the worst they are ever going to be. That's the incredible thing - they only get better from here.
What's inside
Understanding the tools · capabilities & limitations · prompt engineering · content creation · business strategy · tool integrations · advanced techniques · ethics & best practice.
01Understanding ChatGPT and AI tools
ChatGPT is a language model trained on a vast amount of text. It generates human-like text from the prompts you give it, and it's especially useful for content creation, brainstorming and idea generation. The key framing: treat it as an assistant, not a replacement for your expertise. Alongside it I use other tools - image generation, transcription - each for the job it does best.
02Capabilities and limitations
Using it well starts with knowing both sides of the ledger.
Drafting written content, answering questions and explaining, research and summarisation, brainstorming and outlining, analysing text for insight.
A knowledge cutoff, a tendency to "hallucinate" inaccuracies, no real-time information or lived experience, and no memory between conversations.
A friend calls ChatGPT "a very, very fast intern who is very good at their job, but occasionally does acid."
It captures both the efficiency and the need to check the work.
03Effective prompt engineering
Output quality tracks prompt quality - the old "garbage in, garbage out". The patterns that consistently help:
- Be specific and give context - and let it interview you: "Act as a content strategy expert. First, ask me 10 questions about my business, then build the plan."
- Use two-part prompts for iterative work: "Reply in two parts - part one, the updated to-do list; part two, one question to refine it further."
- Specify the output format ("as a table", "as CSV"), assign a role ("act as a sales-page expert"), describe the audience, and name the tone you want.
Then refine. The first answer is a starting point, not the finished article.
04Content creation
The workflow I lean on: ask for an outline, expand specific points, then request multiple title options. The same shape works for social posts ("a 10-slide carousel outline - headline + 2-3 bullets per slide"), email ("15 subject lines under 50 characters"), podcasts ("20 episode ideas with guest suggestions"), and sales pages ("headline, 3-5 benefits, social proof, a strong CTA").
I discard about 95% of what it generates - but that remaining 5% is pure gold. It's a springboard for your own expertise.
05Business strategy
Beyond content, it's a sharp thinking partner: drafting customer-survey questions, building detailed customer personas, walking a SWOT with probing questions per quadrant, outlining a business plan section by section, and brainstorming solutions to a specific challenge ("10 options, conventional to out-of-the-box"). Apply your own judgment to anything that actually matters.
06Integrating ChatGPT with other tools
The real efficiency comes from combinations. Five I use constantly:
Five workflow combos
07Advanced techniques
A few that punch above their weight: two-part prompts for dynamic, conversational refinement; style analysis ("analyse this post's style, then write a new one in it") to keep your voice; simplifying ("explain this as you would to a 15-year-old"); content briefs for writers; and repurposing one post into a thread, an article and a newsletter at once. All of them reward experimentation.
08Ethics and best practice
View these tools as powerful assistants that augment your skills - not replacements for your judgment.
Workshop notes, first published July 2023 - patterns updated for what's stayed true.
Book a Focus Call