Zomalex | Data and AI Training

Gen AI Skills needed

Skills need to flourish in age of AI are not technical but creativity and curiosity

Why creativity matters

AI assistants are general-purpose technology (like electricity) , with an incredibly wide variety of use-cases.

Compare these to other technologies that I teach

Power BI: used for data analysis and reporting
SQL: use for getting data into and out of data stores
Excel: used for everything from shopping lists to financial models (It also has a claim to be general purpose with 1 billion users.)

The capabilities of AI tools are improving at a fair crack although there is argument about exactly how far and how fast. For example, GPT-5 is better than GPT-4

it chooses how hard to think to answer a prompt
it is more accurate, more helpful, more empathetic and less obsequious
but it does not suffer fools gladly

In the best sense, AI assistants are a solution in search of a problem. Our human creativity and imagination will find wonderful use-cases for them.

For example, this right-brain speaker has used AI tools in the last month to:

build this website,
sort out his tax affairs,
asked for advice for a health issue (for a friend of course). The AI has a reassuring bedside manner!
work out where to buy a property,
and finally, forgotten how to Google

The students on my AI course do amazing things in the lab exercises:

speak and chat with their AI assistants, in English, Romanian and Hindi
create poetry in the style of 1940 poets
caption films that only provide spoken English sound track for a Friday night family film night

The capabilities of the AI tools have more advanced features such as

notebooks - on Jane Austen novels, buying a property, university module lectures
story books
study and learn mode
chain of thought / reasoning / researcher modes

Some of these capabilities are immediately and obviously useful. Others need our creativity to find how to put to best use.

Why communication matters

Getting the best from these AI tools depends on good communication. Experts dress up this as “prompt engineering” to sound more impressive.

There is a wealth of advice how to do this (and as model improve, the advice gets updated.) For example:

Be clear, specific and unambiguous.
Provide examples, both good and bad to show the type of response that you want.

Here are the guidelines from our AI course. It does boil down to what we learned (hopefully) in English lessons in school.