You’ve probably heard “AI” and “generative AI” used interchangeably — but they’re not quite the same thing.
AI is the broad umbrella: any system that can perform tasks that normally require human intelligence — recognising images, making recommendations, predicting outcomes.
Generative AI is a specific type of AI that creates new content — text, images, audio, video, code — based on a prompt. Tools like ChatGPT, Google’s Gemini, and Claude are generative AI. So are image tools like Midjourney.
The key difference in practice:
- A traditional AI system might predict which customers are likely to cancel their subscription (a prediction)
- A generative AI tool can write the email you’d send to try to retain them (a creation)
For most businesses, generative AI is the more immediately useful category — it’s the one that helps with first drafts of emails, social posts, product descriptions, meeting summaries, and code.
The tools are improving fast and converging — most major AI assistants now combine generative capabilities with search, data analysis and “agent” features that can take actions on your behalf. The practical question is the same as ever: pick one repetitive task, try a generative AI tool on it, and see how much editing the output actually needs.
How generative AI actually works (without the maths)
Generative AI models are trained on huge amounts of existing text, images or code, learning the patterns and relationships within that data. When you give it a prompt, it predicts — piece by piece — what’s most likely to come next, based on everything it learned. For text, that means predicting the next word over and over until it forms a full response. There’s no database of pre-written answers being looked up; each response is generated fresh, which is why the same prompt can produce slightly different results each time.
What generative AI is good at — and what it isn’t
Strong at:
- Producing a fast first draft of almost any written content
- Summarising long documents, transcripts or threads into key points
- Rewriting content in a different tone, length or format
- Generating variations — multiple subject lines, headline options, or design concepts to choose from
Weaker at:
- Being reliably factually accurate — it can state incorrect information confidently and fluently (“hallucination”)
- Maths and precise calculations, unless paired with a tool that does the actual computation
- Anything requiring up-to-the-minute information, unless the tool has live web access
- Understanding nuance specific to your business that wasn’t in its training data
A simple way to try it
Pick a task you do at least weekly — drafting a client update email, writing a job ad, or summarising a long report. Give a generative AI tool the same brief you’d give a colleague, including context (audience, tone, length). Compare the output to what you’d have written yourself: if it gets you 70% of the way there in a fraction of the time, that’s a task worth keeping AI in the loop for. If it consistently misses the mark, that’s useful information too — not every task suits these tools yet.
