What is AI?

By Daniel Scales

The hottest topic right now: AI. There are lots of definitions about what exactly AI is, and although we could delve into the nitty gritty details of how it internally functions, that’s for another time. Here we are going to reframe the question into what it can do for you. Because, let’s face it, this blog would be too long, and nobody has the time to earn a computer science degree just to figure out AI.

Now, one of the fundamental problems with computers up until recently, is that they do exactly what you tell them to. Although that may sound appealing, it creates a lot of blockers in everything we do.

Ever forgot a word in a sentence: “ Did you remember to send the … erm (insert brain fog here)”, and then proceed to filter through google results for words meaning “send a letter on the internet”? The answer to this specifically is probably no, but you get what we mean. However, with auto predict, the computer will be able to guess what you are talking about. That is a snippet of exactly what AI can do for you and your business, it simulates human intelligence and problem-solving abilities helping us reduce blockers, streamline processes, and explore opportunities we couldn’t before.

What else can it do!?

Although auto predict is great and saves time, it is a small fish in a big pond of problems, so here are some more examples of what AI has already done, because what it could do is great, but you want to know what it can do now!

Fraud Prevention

In 2017 Quantexa used AI to spot fraudulent refund requests on UK trains. Due to the scheme being introduced for refunds on trains delayed for more than 15 minutes. The number of claims were already in the tens of thousands per month but increased 5-fold.

This was a test so exact saving figures were not calculated, but in the top band of 10, the accuracy of the AI system in guessing whether the request was fraudulent or not was 95% with an overall accuracy of 70% from the AI throughout the bands. This was tested on 5,688 peoples accounts. Imagine a similar system for a bank!


Opportunity/ Trend Spotting

A common term which is used a lot is a company being ‘Data Driven.’ Now this is something that is required by a business for efficiency as opposed to making inventory decisions based on which direction the wind is blowing, but what about being ‘AI driven’ as well?

Using AI to initially Identify Correlations and anomalies in data, we can then use human resource to validate these correlations. Imagine having a driver for your business that you were not even aware of because nobody thought to look for it? Or an industry downturn / upturn not being considered when forecasting sales for the next quarter because the analytics team was only looking at your internal YoY data? Customer Segments that nobody had thought of?

The list goes on, but these are just a few ways in which AI can quickly present a massive ROI.


A Polarising subject (do not worry, we know), BUT, when implemented right, chatbots can automate 30% of tasks done by todays contact centre staff!

But Customer support is not all it can do. How about upselling, lead generation and saving abandoned carts?! Now there are lots of figures out there about how Chatbots can increase sales conversions by up to 70%, but it is hard to accurately cross reference all transactions with Chatbot interactions, so here is a list of statistics for chatbot interaction rates for a set of use cases provided by Tidio:

Pretty staggering numbers right? 20% of visitors interacting with an Upselling chatbot?

Being able to be there and know where the customer is in their journey on top of organising an upsell before they leave is a huge value add! Something a human would never be able to do! However, creating the initial lead with a chatbot and following up with a sales representative, is a fantastic way to leverage both human interaction and AI’s ability to be everywhere at once.

Content Generation

Generative AI has already started eating up market share in the creative space. With tools such as Midjourney not only generating art with a few sentences, but with their new consistent character feature allowing you to create multiple images with the same character! Imagine how much time you could have saved if your team were using this instead of drawing each image separately!


Internal Data Assistant

Okay. …so, Imagine Google Search, right? But its AI and it is internal to your business.

It harnesses all the external knowledge of ChatGPT or Copilot, but it does not use your data to train itself, meaning your data is fully secure.
This opens a world of opportunity.

This is not something new either, Microsoft has already integrated Copilot within its Dataverse offerings, allowing you to interrogate your own data!
But why stop with just data?

 Imagine you have a legal case, or a Contract that you are writing up and you want a second opinion. ChatGPT will have it reviewed in seconds. Now you no longer must fully rely on just a second pair of eyes, now you have a computer on your side as well!

How about losing a file and needing to find it, but nobody knows where it is? Ask AI!

Need to create a presentation? Ask AI!

Do you have a set of data that your making decisions off and want to see if you missed anything? Ask AI!

How to get started

Getting started may be easier than you think. For example, Midjourney is a ~$12 subscription which can be started in a matter of minutes, whereas an internal data Interrogator is a bit of a chunkier setup, so it is important to figure out exactly what you want / need.

  • Discovery – Begin by doing an audit of your internal activities and identify opportunities and pain points within the organisation. A fantastic way to do this is by using a ‘naïve expert’ to do an audit on top of an internal audit. This is someone who has no prior experience of the specific business function they are looking at (i.e. find someone external or from a completely different area of the company). This way, they are not indoctrinated into the current ways of working and can instantly point out where things are not working. If you have been in it for too long, you become a part of it!
  • Design – Create a Strategy that is right for your organisation. Harnessing AI is not necessarily an overnight job, as it is still a tool that may need some upskilling and infrastructure foundations to be able to work effectively!
  • Build – Create a plan to implement the AI solution incrementally into the business by doing it in phases. You do not want to implement everything at once and completely change the way the business operates as that will upset the apple cart!
  • Iterate – AI is developing all the time, which means you will need to as well to stay competitive! Keep on top of what is out there reviewing the landscape for even more opportunity!

What if we… just don’t use it?

AI is evolving rapidly, and businesses need to keep up to stay competitive by laying down the foundations to introduce AI into their systems. Even if you are not thinking about using AI right now, there is a big possibility it may be a necessity soon, so being ready is the best way to mitigate risk and future proof your business.

Contact us at DATA³ if you need any help with getting started!