These days, it seems like you can't read a blog post without the mention of AI. The release of Open AI's ChatGPT and DAL-E has fuelled the AI 'craze' to the point that some are concerned it's all hype with no real substance. 
However, AI has been around for years, and cloud vendors such as Microsoft and AWS have been providing some amazing tools that allow people to spin up the specialist hardware and software to create their own AI models and even train them on their own data. 
So if AI has been around for a while anyway, why the sudden spike in interest? To answer that, we first need to think about people's view of AI in general. 
There are essentially three groups that people fall into; 
I know what AI is, I know how my company can use it, and it's really exciting. 
I sort of know what AI is, it sounds very exciting, but I'm not sure how I can use it. 
AI is all hype, I can see no real-world benefits, and it will eventually disappear. 
First, I'd like to address the third group as it can be dangerous to dismiss any new technology. 
In 1895, Lord Kelvin, a prominent British scientist known for his work in thermodynamics (and whom the temperature scale Kelvin is named after) made some notable mispredictions about the future of technology. He said: 
"I can state flatly that heavier than air flying machines are impossible." 
Eight years later, in 1903, the Wright brothers made their historic flight at Kitty Hawk, not just proving him wrong, but kicking off a multi billion dollar industry and a technology we simply could not imagine life without today. 
The release of ChatGPT's GPT 3 and 4 language models and DAL-E image generation 
tools have given the first real glimpse into what AI can actually do for us. However, even these new services that provide a clearer view of what they can do are still being used in a limited way. 
So far, the biggest and most obvious use case for ChatGPT has been the ability to write articles - blogs, letters, emails, even school homework and essays. 
And no, in case you are wondering, this article has NOT been written by AI (a big giveaway is that these articles often end with 'In Conclusion'). 
These examples are extremely powerful and can really help support marketing strategies. However, the models that drive ChatGPT can do so much more for businesses. 
Marketing and Research 
Research is made so much easier and quicker with Open AI. We can use ChatGPT to quickly generate ideas for product development, or DAL-E to generate sample imagery, concepts or even logos. What once took weeks and months for a marketing team to produce can now be accomplished in hours. 
I've already mentioned that the big area of interest for ChatGPT is the ability to generate articles, emails and blogs. However, this is usually actioned through the ChatGPT app - but you can also automate it by programmatically interacting with the AI models. Open AI offers direct access via its APIs, plus you can now create your own Open AI service within Azure as part of their new Open AI component. 
As well as providing the ability to write software that will interact and automate the generation of written information, through the APIs, you can hook up your own data and have that as the source of information. 
This essentially means you can plug in your own data and very easily create useful chatbots for your customers. Again, chatbots tend to be the 'obvious' choice, but the ability to use your own data and automate processes opens up so many more possibilities. 
It can even be used to populate social media channels, such as an automated twitter bot that uses your existing blog to post regularly, or post updates to your Instagram feed. 
Software Development 
Through the ChatGPT console, you can ask it to create c#, javascript or python code to achieve whatever task you have in mind. Ask ChatGPT to “build a simple react application that takes in text data and posts it to a backend API” and the tool will spit out full instructions and line by line code. 
With GitHub Co-pilot, you can take software development to the next level. You can install a plugin to your favorite code tool (for example Visual Studio Code), then enter comments that state what code you need and start creating the function. CoPilot will then provide a complete example of how to write your code - which you simply accept then tweak as required. 
In programming terms, Copilot excels at generating what is called 'boilerplate code'. A big chunk of any software development project is writing code that performs the same or similar tasks - the above GetHtmlResponse is a classic example. 
Therefore, being able to have Copilot do this for you speeds development by a huge amount, and allows your developers to concentrate on building the more complex code. 
Financial Analysis 
AI Models are great at reading through lots of data and summarizing or simply looking for particular aspects. This can come in very useful for understanding the health of a company from quarterly reports. 
With this in mind, one exciting area of research is the use of Open AI to read company reports and quickly understand the key areas that an investor might be interested in. 
For example, we can ask: “please can you describe what the company Microsoft with ticker MSFT does? What are their products and historical performance?”. The AI will then provide a quick summary of the key areas of interest. 
We can then delve deeper asking for insights, competitors and past performance. Finally, we could provide a transcript of the latest results and ask for a meaningful summary. 
In fact, Microsoft and the London Stock Exchange have recently entered into a 10-year partnership to explore this area further. And they aren’t the only ones. Many big investment banks are also starting to build out models to help them better predict the success or failure of investments. 
Although CV scanners have been in use for some time, they are arguably not fair or can miss important aspects. Therefore, an AI tool that can highlight where a prospect’s experience matches a job specification can greatly speed up the review process. 
Rather than having a program simply accept or discount a CV, AI can assist a recruiter in finding the right information and therefore make a more informed choice. 
Conversely, we can also use AI to help us write better CVs. Again, using AI we can input a CV and a job spec and ask the program to tweak the CV to showcase relevant areas of experience. 
AI Ethics 
The recruitment example throws up a very important aspect of AI, and one that has got people worried (ignoring the fear of a Terminator / iRobot style apocalypse of course). 
We mentioned that some companies have been known to employ CV scanners that will automatically accept or reject a CV based on a particular algorithm. An AI program could do the same, and would arguably be better, or at least more accurate, than a human defined set of rules. 
However, it still raises many concerns – for example, how do we know the data the AI was trained on was fair? Because that’s how AI works, we train it on big data sets, and have it make recommendations accordingly. 
But an early attempt at this had an AI bot showing clear prejudice because the data it was trained on was prejudiced! This is clearly unacceptable. 
A better use, therefore, is as I state – use the AI to highlight areas of interest based on a job spec, but still hand the result over to a human. In this way AI becomes a tool, it isn’t making decisions for us, it is just helping us make them ourselves. 
In conclusion (just kidding) 
We've only just scratched the surface of what we can use AI for. These are just a few examples that I have come across either through my work with our clients or through personal interests. 
It's also worth noting that AI is such a flexible technology that the most powerful use might not have been thought of yet. 
For example, ARPANET was originally a computer network that was designed by the US Department of Defense as a resilient network that could withstand a nuclear attack - today that network has grown worldwide and is known as the internet. I suspect AI will also power an as yet unknown technology. 
Over the next few weeks we will explore in more detail different use cases in partnership with some of our amazing customers - so watch this space. 
If you would like to learn more, or would be keen to see how AI could help your business, reach out to Brett Hargreaves at 
Share this post:

Leave a comment: 

Our site uses cookies. For more information, see our cookie policy. Accept cookies and close
Reject cookies Manage settings