How to actually
use AI in your
business

What is AI?

AI, or Artificial Intelligence, has actually been around for ages - it was a term first coined in the 1950s - but there’s been a real, rapid pace of change over the last decade or so. It will have come to most people’s attention in the last 12 months or so since the launch of ChatGPT.

ChatGPT - a novelty, or good for business?

ChatGPT is an AI chatbot, using what’s known as a natural language processing model.

Simply, you can ask it questions in human text and expect a human-like response back. It’s not like a Google search which will come back with a list of terms; it understands the question and is contextual, so it can understand what you’re trying to say and consider follow-up questions to narrow the results. It understands its environment and what is going on.

Most people will have heard of ChatGPT; some people will be using it; a smaller group of people will be using it in their business.

If you’re in the middle group and like using ChatGPT, then why aren’t you already exploring AI in your business? It might seem new, but AI can help solve your problems.

Examples of AI in common use

Let’s consider the example of Google Translate, which typically has the reputation of being a bit ropey - you can tell it’s not using natural language processing. If you were to use ChatGPT for the same translation, it should do a better job, as it understands meaning and is intuitive in the same way as a human. Intuitiveness kicks in when you have to make that contextual decision.

Another example: you need to get to work on time. There’s a traffic issue on the road ahead of you. Using your intuition, you know that if you take the small back road coming up, you can probably skip the queue.

But AI could give you a better choice. How?

Well, for starters, it possibly has access to more data than you have. You can only work off historical knowledge and the sight of what is in front of you (because you wouldn’t be using your phone to Google traffic conditions whilst at the wheel…). Whereas, from the location of smartphones and in-car computers, AI knows loads of people have already turned down the side road, there’s a lorry blocking the path, and you have no chance of getting through. So AI might suggest waiting for the road to clear or trying a different detour.

There’s no getting around it - AI with a good set of data behind it is a clever beast!

Training AI is the same as training your dog

Your dog doesn’t know anything until it’s trained. You ask it to sit, it doesn’t know what to do.

Repeat it, give it a nudge down, give it a treat - the puppy starts to learn the correct behaviour.

The concept is the same with AI - you reward it and penalise it with points when it does a good or bad job.

Both dog and AI gradually build up a picture of how they should behave in that environment.

Great, but how does that fit within business?

The right type of AI technology for your business depends on the problem you are trying to solve.

Not all AI comes in the form of natural language processing like ChatGPT. There’s also:

• Machine learning (entry-level AI, learning about data)
• Deep learning (like speech recognition, autonomous driving)
• Computer vision
• Reinforcement learning (e.g. is this a dog? What about this? And this?)
• Robotics

…and more. There are lots of types of AI. So instead of figuring out what type of AI you want to implement in your business, be solution-led and think about the results of what you want AI to achieve.

For example, AI might be used to:

• Improve customer service, with a chatbot that you’ve fed with all your FAQs and knowledgebase. It can answer customer questions pretty effectively and free up human time for bigger, more complex issues.
• Ensure that personalised campaigns reach the right people at the right time for your sales and marketing teams, based on known data on conversions.
• Apply computer vision to a production line for circuit boards. AI knows what a healthy-looking circuit board should look like, so it can scan for defects and detect when soldering has gone wrong as it scans the end of the production line - then refer it to a human in case of a problem.
• Detect fraud and assist with cybersecurity within financial services, as AI is so good at parsing data and drawing conclusions, finding those anomalies.
• Save lives in healthcare. Blood test results can be run through an AI trained on the right knowledge and tell you what actions you need to take to improve your health.

Within industry, AI can be used across a wide range of applications from supply chain optimisations to robotics and automation, process monitoring to energy management. Google, for example, is using AI in their data centres to monitor energy usage when cooling their big processors, therefore cutting down on their carbon consumption - a useful tool if looking to reach net zero. It can even be applied to things like worker safety, with a camera monitoring those moving into hazardous areas and warning people if they’re not wearing the correct PPE.

In short, AI can be used in your business right now

This isn’t Tomorrow’s World stuff. AI is here now and ready to be used.

The bad news is that it can take a bit of time, energy, investment - and the essential: data. But it is possible, no matter the size of your business.

So how can you get started?

Step 1: Define the problem. Whether you have big problems or small ones, start with the smallest problem first. This’ll help you prototype and ensure you get the right ROI when you move onto the bigger challenge.

Step 2: Understand the problem. Speak to the people involved. Understand the root causes and patterns within the problem.

Step 3: Identify potential AI and software solutions. If you have data all over the place or within different off-the-shelf products, consider firstly whether custom software can help you bring it all together and make the introduction of AI easier.

Step 4: Data pre-processing. Get hold of historical data sets. Tidy it. Make sure it’s clean and comprehensive. The better the data, the better the result from AI. Remember, if you’re teaching a dog to sit and give it a treat when it jumps, you won’t get such a good result!

Step 5: Develop and integrate your AI. And test! That’s key.

Step 6: Iterate and improve. Keep testing. Add up-to-date data. Business problems aren’t static and change all the time, so you’ll need to keep investing in and training your AI.

So, what are you waiting for?

Like the pagers, PalmPilots, PCs, laptops, Blackberries, and smartphones which have all come before, businesses will need AI at some stage. Other people and your competitors will be using it, and if you’re not thinking about it for your business, you’re doing yourself a disservice.

So start the cogs in motion, understand your problems, and gather your data.

At Switchplane, we’re thrilled to be able to bring the latest AI solutions to every business, no matter how large or small. It all ties in with our ethos of creating meaningful change. Wherever you are on your AI journey, get in touch with us for a free consultation on how you can use AI within your business - and make that AI office dog a reality!

At Eastbourne DigiFest in September 2023, Switchplane MD and CTO Tom Fielder gave his TedTalk-esque presentation on why AI is a bit like your dog and how to actually use it in your business. Watch it in full below.