Chapter 3 - 31.03.2026 - AI

I’m Realising Most People Barely Use AI Properly

The more I use ChatGPT, the more I realise most people are massively underusing it.

A lot of people still use it like a better search bar or a place to clean up an email. That is fine, but it is surface-level use. It is like owning a full toolbox and only ever reaching for one small screwdriver. The tool can do far more than most people realise, but only if you stop treating it like a novelty and start treating it like infrastructure.

That has been the biggest shift for me.

I no longer really see AI as just something that answers questions. I see it more as a working layer that can sit beside me and remove friction. It helps me organise information, write faster, challenge my thinking, compare documents, structure projects, and cut down time spent on repetitive tasks that do not deserve too much mental energy.

What makes it powerful is not just intelligence. It is the combination of memory, structure, speed, and flexibility. It can act like a drafter, researcher, organiser, teacher, second set of eyes, and sometimes even a pressure-test for my own thinking. The real advantage is not in asking clever one-off questions. The real advantage is learning how to build it into the way you already operate.

One of the first things I realised is that ChatGPT becomes more useful the more context it has. If every chat starts from zero, then you waste time re-explaining yourself, your work, your tone, and what you actually need. Once you give it that context, the quality improves. It starts feeling less like a blank machine and more like something that already understands how you think.

That matters in practical ways. For example, I use it to help draft professional communication, tighten operational emails, organise ideas, and turn rough notes into something more structured. Instead of starting with a blank page, I can start with direction. That reduces drag immediately. It is not that AI replaces the thinking. It helps remove the slow, repetitive part between thought and output.

Another major shift is using it for voice rather than only typing. That sounds simple, but it changes the way you work. When you are walking, driving, thinking, or just mentally overloaded, speaking is often faster and more natural than trying to type perfectly. Sometimes the fastest way to get clarity is to talk through a problem in rough form and let AI structure it after. That is useful when your head is full and the issue is not lack of ideas, but too many at once.

I have found that one of the strongest uses of AI is as a processing tool. Not in the dramatic sense people talk about online, but in a very practical one. If there is a large amount of information to get through, AI can help cut the first pass down quickly. Documents, notes, screenshots, reports, long transcripts, policies, proposals, comparisons, planning materials. Instead of manually trying to sift everything line by line with no system, I can use AI to extract the signal, summarise the core points, highlight differences, identify patterns, or help me turn raw input into something usable.

That is where time really gets saved.

It is also where most people still do not go far enough. They ask one question, get one answer, and stop there. But the better use is to turn AI into a workbench. Give it context, documents, constraints, examples, goals, and then keep refining. That is when it becomes genuinely useful.

Projects are another area where this becomes obvious. Separating work into clear containers changes the quality of output. Instead of mixing everything together, you can create dedicated spaces around specific goals. A work project. A writing project. A business idea. A health plan. A property issue. A training structure. A presentation. When all the relevant material lives in the same place, the results become more consistent and more specific. That matters because not every part of life should share the same instructions, tone, files, or priorities.

There is also a very underrated use of AI that I think more people need to adopt: using it against yourself.

Most people want AI to validate them. They want it to agree, support, confirm, and reassure. That is probably one of the weakest ways to use it. A much better use is to make it challenge you. Ask it where the gaps are. Ask it what you are not seeing. Ask it how your plan fails. Ask it to argue from the position of a critic, competitor, investor, sceptic, or operator. That is where it becomes genuinely valuable.

I have found that helpful because people are often too close to their own thinking. Once you build an idea, you naturally want it to be right. AI is useful when it interrupts that bias. Not because it is always correct, but because it forces pressure onto your logic. Pressure reveals weak points faster than praise does.

Then there is the learning side.

A lot of people now use AI to sound informed without actually learning anything. That is a trap. It gives the appearance of competence without the substance. Used properly, though, AI can speed up genuine understanding. The key is to make it teach in layers. Strip out the jargon. Explain things simply first. Then increase complexity gradually. Use metaphors. Use examples. Ask it to test you. Ask it to correct your understanding. Ask it to explain the same thing three different ways until it clicks.

That is far more useful than copying polished explanations you do not fully understand.

I also think the visual side of AI is underrated. People focus too much on text and forget that image, layout, and screenshot analysis can be just as useful. You can upload an image of a dashboard, a layout, a room, a diagram, a product, a damaged item, a sketch, or even a spreadsheet and ask for direct analysis. What is wrong with this? What stands out? What is the one thing to fix first? That can save time because sometimes the problem is visible before it is verbal.

The same applies to comparing information. AI is useful for spotting changes, contradictions, inconsistencies, or improvements between two versions of something. Contracts, policies, reports, letters, drafts, operating procedures, proposals. It can help show what changed and why it matters. That is a good example of where AI works best as a second layer of inspection rather than a replacement for judgment.

There is also a broader point here about automation. Not everything needs your full attention. A lot of daily work is repetitive, administrative, or structurally predictable. Reminders. Recurring prompts. summaries. Routine checks. Organising lists. Rewriting drafts. Pulling together inputs. Pre-structuring outputs. The more of that you can hand off, the more mental space you preserve for actual decision-making.

That is where I think AI starts to become genuinely strategic.

The point is not just doing more in less time. The point is reducing wasted effort. It is about protecting your attention from low-value repetition and using that recovered energy where it actually matters. Good judgment. Hard decisions. Creativity. Relationships. Leadership. Real-world execution. AI should be removing friction, not creating dependency.

That is probably the balance that matters most.

I do not think the goal is to let AI think for you. I think the goal is to let it handle the parts that slow your thinking down. There is a difference. Used badly, AI makes people lazy, shallow, and overconfident. Used properly, it makes people faster, clearer, and more deliberate.

For me, the biggest benefit has been in three areas: structure, speed, and clarity.

Structure, because it helps take rough thoughts and turn them into organised output.
Speed, because it reduces the time between idea and execution.
Clarity, because it helps cut through noise and expose what actually matters.

And the more I use it, the more obvious it becomes that most people are still only scratching the surface. They are using one function when there are twenty others sitting unused. They are asking it to answer questions when they could be using it to manage context, challenge decisions, teach concepts, analyse visuals, compare information, support projects, and automate repeated work.

That is why I think AI is such an advantage right now.

Not because it is magic.
Not because it replaces people.
But because most people still have not learned how to use it properly.

The edge is not just access to the tool.
The edge is knowing how to think with it.

And once you start doing that, it stops feeling like a chatbot and starts feeling like a serious operating advantage.