Matt Crick, Senior Systems & Training Consultant in our Business Solutions team gives us an introduction to AI, what’s brought it to everyone’s attention, and why we should care?
What is Artificial Intelligence?
Artificial intelligence, or AI, is the ability of machines to do things that normally require human intelligence, such as understanding language, recognising images, making decisions, and learning from data. AI is not a single “thing” at all, but a collection of methods and tools that can be applied to the various problems and domains that we have. They have been around for a while now and you will already be familiar with, and using them:
- Machine learning: Predictive models based on data and statistics – used for predicting the weather.
- Anomaly detection: Systems that detect unusual patterns or events, enabling pre-emptive action – found in security and monitoring systems.
- Computer vision: Applications that interpret visual input from cameras, images, or videos – often used in KYC processes and in speed cameras.
- Conversational AI: AI agents, (or bots), that can engage in dialogs with human users – ChatGPT anyone (or that ghastly paperclip from ye olde Office applications)?
It’s in this last area that we have seen AI recently get pushed into the limelight in the past few months. With conversational based apps that feel like we are talking to an intelligent creature and getting unique outputs.
“a highly detail picture of a futuristic circuit board with a brain in the middle of it. hyperealistic, 4k render, HDR, god rays, ray tracing.”
What is all the fuss about?
ChatGPT took the world by storm with its scarily impressive abilities to not only understand language but to also create and produce things such as stories, poems and computer code. Being able to create (often) new things from a set of natural language instructions is what we know as generative AI and it couldn’t be done without the use of Large Language Models (LLMs).
LLMs are capable of understanding and generating human-like text by processing vast amounts of data. They are a key component of generative AI.
It was this that got so many people using the technology – the ability to be able to create seemingly new creations from simply describing them. Suddenly everyone could feel like a video artist, a poet, a musician or a highly skilled visual artist, and it truly does feel like magic.
Wherever you sit on the ethical concerns and conversations to be had around this technology there is no denying, it can do some very impressive stuff and is improving at doing so at an alarming rate.
Go try some applications out if you haven’t already – a few I have used and been impressed by are: ChatGPT, Sudo, Microsoft Copilot and DALL-E*.
*(I have used DALL-E to generate the images in this article with the prompts underneath. If you use the same prompts, you will get a totally new picture based upon the same idea! Am I graphic designer now? Is DALL-E?)
“a monkey looking at a fire fairy”
What does this mean to me?
At the very least you are going to be continually using products and services that utilise AI to provide you with a better experience, and you may not even notice. At best you will arm yourself with the correct AI tools to improve your life, work, productivity and creativity.
It is in the best interests of professionals and businesses (and all people really!) to capitalise on the benefits that can be had from using AI and integrating AI tools, features and services into existing processes and solutions.
Let me to take you through some of the ways AI has helped me to help give a broader picture.
- Music Production – there are some great applications of AI in the audio plugin world. From smart compressors and EQs(equalisers) to sound separation tools that allow you to, for example, isolate a vocal from a full track or reduce the background noise on an audio sample.
- Writing – generative AI can be super useful in generating, writing, critiquing and summarising content. It has helped with a few ideas for this piece (although I prefer to largely write my own material to retain my voice). I go to it for various things such as E-mails, Power Point slides, articles and documentation. It’s great at coming up with a plan to cover pretty much any subject you can throw at it, which can then be used as inspiration for what you are working on.
- Googling – or should I say “Generative AI’ing”. I’ll often just ask Copilot (more on that later) for queries I have. Rather than me going to a search engine, looking at a few of the most likely relevant results, skim reading each, gaining an understanding of the material and coming up with a summarised idea of it in my head… I just let Copilot do that for me, in a few seconds.
- For example, “Tell me the top 5 things people do in New York with a total budget of $1000 in October. Don’t include anything that includes big heights.”
- Coding – many a times have I asked Copilot (Microsoft’s generative AI tool) to write me a quick bit of JavaScript or to talk me through more complicated C# coding solutions. Not only is it a great at teaching these things but also coming up with code solutions. Of course, anything it produces is highly scrutinised, as it does gets things wrong on occasion!
- Enhancing Client Offerings – being able to access many AI services and tools means I can now offer things to my clients that were previously outside my skillset.
You want automatic translations of any document stored in a particular folder into 5 different languages? No problem!
You want to automatically create an expense to be reviewed in your platform of choice when an employee e-mails in a receipt? I got you!
You have a large datastore and want to index it so you can query it with natural language (“tell me all documents that were created between 2011 and 2014 that mention the company “Stellar Optics” in a negative way”)? Hold my beer.
“make me a realistic picture of an alligator in a puddle. The puddle should be in the middle of a floating land mass that is contained with a spherical dome. In the background it is sunset made of intense oranges and reds with the silhouette of flying birds.”
How do I get such magic?
AI Apps are springing up at a crazy rate – it’s the latest gold rush. One of my primary concerns when using the technology is how is my use of it being used to train it and other AI applications. On a professional level this question is even more pronounced as there are legal data requirements to consider. This is why I prefer staying within the confines of Microsoft infrastructure.
Everything you do with AI within your Microsoft tenant is private to you. All the data used and submitted to any aspect of AI remains yours and is not used to train any other models or applications. This means I can play around with AI, ask it questions about my data, get it to check over my code and to generate me content without having to worry about all the data involved in doing all that.
Microsoft offers some great products and services for consuming and using AI features and solutions, most of them are not free (but considering the security and governance they offer it’s usually a small price to pay!). Each could be its own article, but I will try to briefly go over them.
M365 Copilot – an AI agent that you can find in all your Office applications and an ever-growing number of other M365 applications. The best thing about this use of AI is it has access to all the data that you have access to (e.g. is security aware) within your tenant. So, you can ask it to summarise your week and it will include information from files you have worked on, meetings you have had, e-mails you have received etc. You can even ask it to find an hour in an afternoon sometime next week where you and specified other colleagues are all available and it will quickly give you some options! Of course, it can also generate content in any app (Word, Excel, PowerPoint etc), you just need to be able to describe what you want, it’s amazing.
Copilot Studio – a quick way to produce chatbots and agents to handle specific tasks like helping people with the information contained in your website or a particular set or documents such as your HR policies. It can even handle complex interaction with external systems, the sky is (not even) the limit.
Azure Cognitive Services – a suite of services that developers can use to build AI solutions without having to deal with the underlying algorithms and models. Cognitive Services provide pre-built and pre-trained functionalities for common AI tasks. It includes some very handy services:
- Open AI Services – build your own copilot and generative AI applications with cutting-edge language and vision models.
- Search – be able to quiz and index your own data that is hosted outside of Office.
- Content Safety – automatically detect offensive or inappropriate content in images and text.
- Translator – translate any text, speech or audio into another language.
- Speech – use a pre-trained AI voice (or train your own!) to talk to users or generate audio using provided text or generated content.
- Vision – read text, analyse images, and detect faces with optical character recognition (OCR) and machine learning.
- Language – build conversational interfaces, summarize documents, and analyse text using prebuilt AI-powered features.
- Document Intelligence – apply advanced machine learning to extract text, key-value pairs, tables, and structures from documents.
How are organisations using AI?
Artificial intelligence is a game-changing technology that can help transform your businesses and gain a competitive edge. There are many ways this could manifest within your own company, one of the most common ways is by saving you time.
Easy AI wins for a business usually come by identifying processes and operations that can be improved by implementing one of the AI offerings mentioned above. It is thinking about all the things AI can do (see, hear, understand language, identify things, summarise, etc.), and how quick it can do it, to figure out how you could piece these things together to help within your own organisational context.
You could allow employees to submit invoices and receipts for expenses via e-mail or a chat channel. Text could be extracted from the receipt and categorised, this data could then be written to your software that deals with expenses to be approved. No one had to key in any information.
How about you have your Teams meetings recorded, then have the audio analysed to get a transcript of what was said whilst recognising who the speaker as to create a set of minutes, which you could then have translated to several other languages along with generating a summary of the meeting. Then, to help with inclusivity, automatically take those summaries and minutes and convert them into audio for those who have visual impairments (and do this for every language).
Companies also have a lot of siloed data – maybe because it’s on-premise, or a lonely SQL server somewhere in the cloud, or a third party holds the data and you can only get to it via an API. There are various ways you can ingest and interact with that data and by using LLMs you can ask questions about it as if you were asking a person. Maybe you’d like to know “how many documents are there that reference our product HJ334 that were created between the 1st April and 24th June and that also contain website links to the product”. Or even “given our financial data what would be a realistic forecast for our IT department next year, explain your answer”.
Anything else?!?
Oh so much more! But I’ve already gone on longer than I wanted in this “small overview” of the subject.
AI is here – it’s not going anywhere. So best figure out how best to use it. Give me a shout or give us call at Prosperity 24/7 if you want to have a quick chat about how we can help you on that journey by being your guide and partner.
I’ll leave you with one last picture and leave you to interpret it.
“could you please make me a picture of what the world will look like in 10 years time given the advancement of AI and likely human actions. Be brutally honest.”
