How to future-proof your business with AI

AI and its capabilities have hit many of the news headlines in recent times, but it’s the release of large language models particularly which has made people sit up and look. The ability to create content from some simple phrases and questions is now unparalleled.

'Passing trend' or 'game-changer' ?

The world of ‘prompt engineering’ is growing exponentially, and the ability to ask the right questions in the right way is already a small industry, in and of itself. This may well grow into its own niche within IT and the larger information, SEO and media industries.

Think Digitally 's Technical Director, Mark Mailer , shares his thoughts on AI's trajectory:

"Whilst the initial shock and awe has somewhat diminished, I believe AI is here to stay – well, large language models and generative AI at least. I do not believe that the current range of publicly available tools are true AI in that they are ‘thinking machines’. However, that doesn’t diminish their impact and ability to help deliver transformative results.

I believe that AI, and Chat GPT in particular, has somewhat dropped from the spotlight because it is still difficult to implement so that it makes significant headway towards meaningful commercial output. When it became accessible to the public, everyone was thinking - “great, I’ll just get AI to do this, that and the other!” - but the engineering to integrate an AI system is complex and requires a lot of time and effort. So, it’s not the panacea many a boardroom was hoping for."

Generative AI and software development

AI can be a means to help condense and refine thoughts into either verbose or concise statements and documents. By stating the problem you have, and an overview of the potential solution as you see it, Chat GPT can do a good ‘first pass’.

Then, by continuing to ask further questions and add more points of clarification, you can create a very concise project brief and requirements document. You may still need to enhance with industry-specific knowledge, or flow charts and graphs. However, some plugins can help with these too.

Being able to validate code on-the-fly, or write a statement of what you’d like some code to do, and get a reasonable or excellent block of code is a huge time saver. Mark Mailer takes us through an example:

Try asking ChatGPT the following;

“Please write me a method in PHP which receives a date, and outputs the number of working days, also excluding public holidays, between that date and now”.

Hopefully you can see the result is pretty impressive. Note > I always remain polite and remember my “pleases” and “thank yous”... after all, if AI systems run the world one day, I’d rather be on their good side!

Being able to ask simple questions in the context of your code, and getting a lot of code written in a matter of seconds is amazing and can save so much time.

Possible threats

This is where things can get a bit crazy with speculation...

Yes, there are threats - when the above example becomes so sophisticated that you can ask “Please write me a CRM system and deploy to AWS” then there are a lot of developers, dev ops and engineers potentially out of work.

That said, for the foreseeable future, in my opinion, we will still need a lot of QA, testing, and the ability to understand wider connections, commercial context, business context and how systems and people interact and interlink.

As it stands, Generative AI can write blocks of code and stand-alone statements, but still is someway from creating entire systems, and even further from creating interconnected systems. I’m sure it will happen eventually, but timescales will be the factor I’d be unwilling to speculate.

There’s also the weaponisation of AI. When sophisticated, highly orchestrated attacks can be performed with incredible accuracy, the tools to defend against these will be required to keep up.

Similarly, when advertising can be 100% customised to you, from the language to the video and images displayed, the ability to sway entire populations becomes child's play. All we need to do is reference the work of Cambridge Analytica in both the Trump Presidential Campaign in the USA, and the Brexit Campaign in the UK, to see how highly targeted advertising {based upon a person's history and profile} can have far-reaching consequences.

As these tools become more available and more understood, this space may become the new battleground over social media. Already, there are companies which help to identify ‘deep fakes’ in near real time to fight against propaganda and misinformation.

Potential opportunities

On the positive side, the opportunities are vast and numerous...

15 years ago, if someone said, “I am going to post content on social media and get paid for it”, you would think they were madder than a box of frogs! Nowadays though, this is a credible job for many, and with headlines like - “University to offer Ireland's first degree in 'influencing'” [www.rte.ie/news/education/2023/0925/1407246-influencing-degree] - not only can you get a job using social media, you can also get a qualification!

It's almost Inevitable that the world of AI will create a vast seam of jobs and career opportunities - from prompt engineering, development integration, to data analytics, and directly related professions we haven’t thought of yet. There will also be many indirectly related professions from content writing, imagery and video creation to the performing arts. This will impact search engine optimisation (SEO) and the search industry with what it can consume and index, and what should be consumed and indexed in any meaningful way.

The reality is that I certainly cannot fathom the nature of jobs and industries which will come about from the rise of AI. I’m sure many great minds will be hard at work thinking of what that could be, but like all new technologies, the applications are only limited by our collective imagination.

What happens next?

As more tools integrate AI from code generation and review, and the models are further refined to take nebulous concepts and define them in greater fidelity, the speed and precision of what can be built will continue to increase.

What we currently see as ‘rapid development tools’ could one day look like the Wright brothers’ first attempts to fly, versus the likes of Virgin Galactic and Blue Origin space flights! Our minimum viable product and proofs of concept will be more complex and complete, as the tools continue to develop from simple ideas and base templates.

This will place greater emphasis on the roles of architects and system designers, and the business analysts will be converting ideas into concrete concepts.

Conversely, without a clear idea and direction, the potential to deliver nothing of meaning will be just as great in the future as it is now. The amount of time lost may decrease, but to deliver something meaningful still requires a clear vision of what the outcome should be... and currently AI is not reaching its full potential in that respect.

As a QA and testing tool (a ‘first pass’ so to speak), it should start to yield great results and already IDE’s (Integrated Development Environment) code checking and syntax highlighting helps significantly, especially compared to the days of writing code in notepad.

We have come a long way, and there is much further to go. The future is going to be very interesting!

If you would like to discover more, please reach out to us!

Email: [email protected]

Web: Contact Us

LinkedIn: Mark Mailer

Tel: +44(0)333 012 0099