It feels like AI is everywhere these days. Whether you have tried it or not, chances are it’s been pushed in your face with every tool you try and use, from making websites to email marketing, image generation, product description…the list goes on and on. Apparantly AI will have your job in the near future, and teenagers on tiktok are generating £££’s per month lying in their beds!
Its not really clear where this is all going and there is a feeling you either ride the wave or get crushed beneath it.
The main concern for me is the massive amounts if investment big tech and private equity are pouring into AI and seemingly, unstoppably getting rich off it. But what does it actually mean for us, the small business owners, creatives, and everyday users?
AI Runs on Tokens, Power, and Data
Some odd statements about AI: every single query to ChatGPT (so simply asking a question on ChatGPT) uses the equivalent of a glass of water which is being passed through cooling systems. Articles are popping up daily on how the water used to cool massive data centres could literally cause shortages in water supplies for housholds near these data centres. And then there’s sheer processing power needed – Google is apparently building several nuclear power stations just to keep up with AI’s energy demand.
The Guardian report on potential water shortages for data centres in the UK
Googles Blog Post on the Kairos Nuclear Power Agreement
This raises real questions about the sustainability of the tech we’re being encouraged to rely on. It’s not magic, it’s energy-intensive and resource-hungry.
Chatbots and Agentic AI
AI is already creeping into our daily business tools. We use a chatbot on our yarn shop’s website, it’s decent, but needs ongoing tweaking to keep it helpful. This is what people refer to as “agentic AI” these are tools trained on your data that can act semi-independently.
But the magic is only as good as the upkeep. You still need to feed it, train it, and monitor its usefulness.
Prompting Is the New Literacy
Using AI effectively comes down to how you ask it things your “prompts.” There’s a whole world of courses teaching prompt engineering. Ironically, you can even ask ChatGPT how to better prompt ChatGPT!
It’s like learning to speak a new dialect one where clarity and specificity matter more than ever.
- Free Courses from OpenAI (who are ChatGPT) https://openai.com/global-affairs/scaling-the-openai-academy/
- Linkedin Learning Courses: https://www.linkedin.com/learning/prompt-engineering-and-ai-agents-with-chatgpt/the-new-age-of-prompting
- Scott Galloway’s ‘Section’ – Paid AI Academy: https://www.sectionai.com/ai-academy
Machine Learning (ML) in Simple Terms
Machine learning is just computers recognising patterns from data, then making decisions or predictions based on that. No one is sitting there telling it exactly what to do it’s figuring it out by crunching data.
Think of it as AI’s backbone: the bit that lets it learn, adapt, and improve over time.
AI in Action: Everyday Tools and Tactics
You’ve probably used AI without even thinking about it. Some examples:
- Chatbots: Gorgias, Tidio, ManyChat
- Social media automation: Tools that write and schedule your posts
- SEO support: I use AI to write meta descriptions as it can keep to the character limit and copy what I have already written.
- Copywriting: Jasper, Copy.ai, and of course, ChatGPT. These can help generate product descriptions, blog posts, and emails. You don’t have to use them word-for-word, but they’re a good starting point.
The key is to make AI work with you, not for you.
AI Isn’t Original, And That’s Fine
Generative AI doesn’t create anything new. It’s been trained on vast amounts of data and is just really good at predicting the next word, based on everything it’s learned. It’s imitation, not invention but clever imitation nonetheless.
OpenAI’s tokenizer visual tool is a great way to see this in action. It breaks down language into tokens (common sequences of characters), and models learn how those tokens relate.
The key point here is that if you are using ChatGPT for free then your ‘tokens’ are limited and your data is open to be used. That’s the pay off for um..not paying. If you then subscribe to ChatGPT you get increased privacy and larger number of tokens to play with. Althought its worth noting at the time of writing Claude is providing a huge number of tokens. I foresee this changing fairly soon.
See the ‘Tokeniser’ in action here:
https://platform.openai.com/tokenizer
Email Marketing: Smart Segmentation with AI
Platforms like Omnisend and Klaviyo now bake AI into every corner of their systems. You can use it to:
- Segment your customer lists
- Target specific buying behaviour
- Automate campaign flows
I still write the final email myself, but AI helps me brainstorm subject lines, segment audiences, and automate repetitive bits.
One tip: use tools like NotebookLM (powered by Google Gemini) to create a walled garden of your own notes and data. It helps you stay focused and query only the info you feed in, great for keeping things tidy and private.
Ads: The Big AI Power Play
If you do one thing with AI, do this: link your product catalogues to Meta Advantage+ and Google Performance Max. These tools use AI to constantly optimise ad campaigns.
When it works, it really works. We had a period where our return on ad spend shot up just by using these tools smartly. But over time, things can plateau, and that’s when you need to refine better content, smarter audiences, tighter data.
Still, these ad tools are the most powerful AI use case for small businesses right now.
A Word on Ethics and Data
Don’t go dumping client data into ChatGPT or any LLM without serious thought about privacy and GDPR. Yes, you can pay for upgraded access with better protections, but you’re still dealing with black boxes to some extent.
And let’s not forget these models are trained on human-created content, often without consent. Artists, writers, and creators are understandably raising objections.
https://www.theguardian.com/technology/2025/jun/30/ai-techscape-copyright?CMP=share_btn_url
What About AI Images?
AI-generated images? Something of an anathema for me at the moment. They are fun no doubt, but you can almost always tell, and even if you cant tell, that’s kind of the red flag that is the central problem. In general if its prompt generated, they are often garish and gaudy and over processed. Not saying you should not ever use it, I did see a photo processed to give an illustrated feel for example, which looked great, but this is using it to augment clear source material ( ideally your own) The guidance here for me is that if you are lazily generating stuff you have prompted (text to image) and mindlessly posting it as ‘real‘ this can end up being dull and overbearing and ultimately off putting.
Nothing beats real photography or even hand-drawn illustrations scanned and cleaned up. It’s more authentic and emotionally resonant. If you do use AI images, make it intentional and most importantly say so.
Final Thoughts: Let AI Take the Drudge Work
AI should remove the boring, repetitive admin. It should make data wrangling easier. Decisions can be parsed through AI to help sort out thought processes. For me the goal is to free you up to work on your own creative process.
If you’re a small business or creative not yet using AI to lighten your load, you’re probably missing a trick. I can help you get started.
Let’s use AI to do the heavy lifting, so we can focus on what really matters: the work only humans can do.
Leave a Reply