In the first of our AI webinar series, we explore the concept of AI and deliver a high level overview of what it can do and the sort of things you need to be thinking about if you're curious about starting to use AI in your business.
Alastair Struthers
So thank you very much for coming to our AI introduction Webinar. Basically, this is our first ever, our lunch and learn ones that we're going to be doing over the next couple of months, frankly, just to try and educate people on AI.
Stacey Williams
Let's do it.
Alastair Struthers
What it can do, what’s out there. I presume many of you probably know me, but if you don't, my name is Alastair. I'm the Customer Success Manager for IT Foundations, and I'm joined by Stacey, who is our, frankly, I'm going to describe her as our AI evangelist.
Stacey Williams
Which is a loose term.
Alastair Struthers
Yeah, she's very enthusiastic about AI. There's not much about it she doesn't know. She came to us a couple of months ago, and joined the team. She's in our sales team, but she is very much an AI enthusiast, let's put it that way. So, we thought between the two of us, we could probably cover most of what you need to know, at least at this stage to give you an introduction to AI.
Alastair Struthers
Let’s just dive straight in. Stacey, do you want to jump to the agenda? So this is effectively what we're going to talk about today. We're going to cover what AI is, why AI matters now. We'll talk about the difference between chat AI and agentic AI, which is a phrase you may or may not have heard, but it's effectively what's coming. It's kind of already here.
We'll have a quick look at some of the tools that exist out there already and the platforms that you can use. We'll talk about some industry-specific type examples. I'm not going to go into great detail on that, more just to talk about the fact that many industry tools are pulling AI into their own platforms. And then we'll have a little look at kind of what it can do. From a leadership point of view, as leaders of businesses, what you can maybe get out of AI. Then we'll have a quick, very quick chat about readiness of it all, how do you get your company in a position to start using AI, and then we'll have a little Q&A at the end just in case there's anything we've not covered that you may want to cover. We'll also in subsequent webinars cover in greater detail getting AI ready.
So getting your business ready for putting in AI, it's not as simple as you might think. We'll go into a lot more detail there in the next one, and then we'll go into a bit of a deep dive on Microsoft's Copilot, which is probably the main AI that most companies are probably going to adopt because it's so firmly baked into Microsoft that we may as well learn how to use it. It's there already.
And it can do some really cool things with proper licencing on top. And then the final one, we will talk about how you actually go about rolling out AI, how you actually make it work in your business. So that is all to come. But I think it's probably time I hand over to Stacey to basically to talk about what AI is.
Stacey Williams
Thanks, Alastair. So AI, we all have heard about it via ChatGPT, newspaper articles, scary horror stories, generally speaking. But obviously, it is a technology that has come out. It's been being released for several years now, and it's at the point where it's available for the average person, which is super exciting.
It is essentially just software systems that can recognise patterns, generate content, answer questions, summarise information, and support decisions based on data and prompts. Modern AI isn't magic or human-like reasoning in every situation. Rather, it's a set of tools trained on large amounts of information to perform useful tasks quickly.
So the most visible forms of AI today include language models, image generators, transcription tools, things that you're probably using already potentially in your business, largely around things like Fireflies note taker. Even the note taker that is currently recording this webinar also has an element of AI to it.
For business audiences, AI should be understood as a practical capability layer that can improve speed, consistency, and access to knowledge. It's really good at drafting and summarising content, so emails, reports, proposals. It's good at finding patterns, supporting routine decisions or processes, and making specialist tools easier through the use of natural language. So, for example, instead of using code to build a website, you can use English language directly with these tools, which is pretty cool.
Alastair Struthers
I've put the slide in. This is a totally made-up graph. I'm not going to lie, this is not based on any actual real data, but it does prove a point. It illustrates it really nicely. AI is just the latest in a series of technological jumps that have transformed business. The very first one (I suppose the very first one was Industrial Revolution but if we ignore that), the first one I want to talk about is computers. Computers came along and all of a sudden, we could do things so much faster. We got spreadsheets and we got word processing and all of a sudden, the things that we were doing were significantly easier and faster than they were before and everyone became much more productive. And then somebody came along, a man called Tim Berners-Lee, he went, hey, I invented something called an internet and everyone, hey, this is great.
It transformed the way we do business again. It suddenly meant that we could send emails instead of writing letters. It meant that we could set up online shops and people could do e-commerce. And there are all kinds of things that came with the internet. And that then enabled the cloud, which meant that everything we did stopped being on our local servers and it all went up into the cloud, which was fantastic. It meant that we didn't need to spend as much money
running hardware and we're all much more productive because you could work from anywhere and it was great. And now we've got AI. And overstating the importance and impact of AI is pretty much impossible, I think, at the moment. AI is just transforming the way everybody is working already. And that's even without most businesses actually knowing what they can do with it. I can pretty much guarantee people in your businesses are already using AI, whether you know it or not. They'll be asking chat GPG questions or they'll be asking Copilot Chat questions. They'll be learning prompts and getting suggestions. It's just being used everywhere. And the companies that are going to adopt this fast are the ones who are really going to see the biggest benefit. I have seen so many projections that I've lost track of who's saying what. But I think generally speaking that the consensus from the last one I saw from the government was that they expect a 40% increase in efficiency in the next two years from companies adopting AI. 40% is huge. I mean, that's not to be sniffed at. We'll talk a little bit later on about what that means, as to whether
that's staff becoming less useful in the business, which is not what we think is going to happen, or whether it's staff becoming more productive in the business, which is what we think is going to happen. We'll cover that in due course. We should probably start, though, just by talking about what the general AIs are at the moment, what the chat AI is that people are already using probably in your business, so I'll hand over to Stacey to have a quick chat about them.
Stacey Williams
So you will already probably be familiar with the name ChatGPT. It's become very household and it has largely been the first one that people have taken to in terms of using it in a conversational way. And it's a great general purpose assistant and it's reasonably suitable for everyday work. So if you want to, I mean, it has become the Google of AI for all intents and purposes. You ask it a question, it gives you an answer. You can now have a conversation with it.
It’s a bit like you are talking to someone on the phone. If you're using your phone and you switch on the mic, or if you're using a laptop and switch on the mic, you can chat to and it will respond like you're having a chat. It's kind of cool, but also a bit strange. The next one has kind of... boomed, if you will, it’s Claude, largely because it released Claude Cowork, which changed the game completely, but also that they've now released called Claude Code. And there are many businesses that are exclusively writing code with Code and using their developers to check that code and to sort of proof it.
It is known for its stronger long form writing and document interaction, but it's also brilliant for automation, connecting with different systems. So, for example, Claude connects really well with things like HubSpot, and suddenly you can use Claude to automate a huge amount of your processes. The next one, which we've already briefly touched on is Copilot, which is sitting inside the Microsoft ecosystem, and it helps within documents, spreadsheets, presentations, and it's particularly relevant where organisations are using Microsoft tools. And then finally, we've kind of got the others. The others are your Grok, Gemini, any sort of other tool that falls outside of the ChatGPT code and Copilot group.
They tend to be a bit more fast moving. They can be a bit more janky, but they are very suited to their context. So a really good example here is Grok, which is specifically used within the context and audience of X or Twitter, as it was formerly known. And in many ways, those sort of other AI’s tend to be a
bit more rogue as well, as I'm sure you may have heard about in different articles specifically pertaining to Grok.
The AI platform services do a variety of things. We know that they create, they take your prompts and they create information. They create draught decks, they create proposals and emails. There are plenty of people creating CVs with AI, as we know, but there's also plenty of people creating websites and proposals. It reduces a lot of that blank page time. It gives you faster starts and better assets, which is really useful. It also captures data, so it will transcribe meetings and take surface actions. It will summarise decisions and create searchable records.
It gives you an opportunity for clear follow up and also faster synthesis, which means people can pay much more attention during meetings rather than hurrying to take notes. Though there is some evidence that taking notes is actually best for you, but we won't get into that right now. The workflow piece is also super interesting and again, has become much more prevalent with the release of Claude Cowork.
The options to route tasks, draught updates, create whole sort of workflows is now possible using AI. As opposed to when we used to follow “if this, if this, then that” logic. AI just lets you type what you want and helps you build it, which is a pretty big step forward in productivity.
It gives you that operational support and it gives you routine outputs, which is great. So another really great example here is, I personally use Claude to run a search every Monday for land available for sale in very specific areas. And so every Monday without fail, I get a report back from Claude telling me what new property has come available in the brackets and parameters that I've requested. And I mean, that's pretty cool. That's super basic, but it is a very interesting use of technology to perform these routine outputs. And the final piece is specialist assistance. Increasingly, we're seeing AI become more embedded in profession specific software. One example, not necessarily profession specific, HubSpot has its own AI tooling sitting inside the platform. As much as you connect it with external tools, the internal AI is also really well suited to carry out specific tasks within the platform itself.
It gives you intel, guidance, and in many cases sector specific context, which is great leverage to have.
So talking a little bit now about some of the SaaS tools that are leveraging AI at the moment. Canva is a great example. Suddenly you can use AI to make your creative production much quicker and much more accessible. Teams can generate visuals, draught copy, build presentations, resize assets, all by using the AI tooling.
It makes for much quicker in-house design, without necessarily needing to go through the process of recruiting a full design person working for you.
The next piece is Lovable, which is a personal favourite of mine. It's a website builder, but it actually is way more than that.
It builds presentations and assets and social media kits. It's a very cool, fun tool to use, but it represents a new generation of AI assisted product and app creation tools. As a non-technical person myself, I can just say what I want and it will just build it. It's very fun to use, but essentially it is for building workflows, rapid prototyping. It makes changes really quickly and I have to say the credit usage is very generous in my personal opinion.
The final one here to note is Fireflies. There are so many AI notetakers out there that are very similar, but they will ultimately record the meeting, transcribe it, summarise the meeting notes, what happened in the meeting, the actions, each person's talking points, and it will assign the actions to the people with the caveat that if you're in a room with multiple people, that becomes a bit trickier. But generally speaking, these tools are really smart. Some are better than others with the level of detail they go into, but generally they all work really well for those meeting workflows and action captures.
Alastair Struthers
Let me just jump in for a sec, Stacey. You mentioned it, something we haven't actually touched in here. We maybe should do. You mentioned the credit usage for some of the tools. Could you, quickly explain how that works, what that actually means, if people aren't familiar with it.
Stacey Williams
Well, yeah, I think it's ultimately black magic, but what it is at the surface level is, so with Lovable as an example I can describe is, you know, I pay for an account to do whatever I want to do, and it will issue me a certain amount of tokens that I can then use against my projects.
For every prompt I give the AI tool, it will deduct credits. Now, it's not one for one. This is why I say it's black magic. It's really not clear, but another tool, clay.com, which is largely for sales and lead generation and things like that, that one will give you, say, one enhanced contacts industry for 0.1 credits. It is really random across the different platforms, but this principle of credits is what all of them are using. You will see things like if you're using ChatGPT for free, it will get to a point where it will say, you've run out of credits for today, come back after 4 P.m. and try again. It's a really good way to make you buy the products, but it's also a way for them to manage, I suppose, the loading at their end in terms of how many requests are going through and things like that, and obviously to monetize the software. So credits are a thing. Obviously, the more credits you have, the more you can do.
And as we're seeing more and more frequently, a lot of these tools are very quietly pushing updates to require you to buy more credits in order to use the system at the same levels you have been using it. And they're all doing this. It's very annoying. But things like Fireflies tend to have this problem a lot less because you are just paying for you know, X amount of meetings per month, it will turn out the transcripts and notes and actions and things, and that's fine. But your Canva, your Lovable, your ChatGPT are all much more prompt-based and therefore using credits in this way. Hopefully that has explained somewhat, and I am not saying that that is all correct according to the gospel, but that is basically my summary of how credits work.
Alastair Struthers
I think that's pretty accurate.
Stacey Williams
Lovely. OK, good. Moving swiftly on.
Alastair Struthers
We mentioned earlier on that a lot of these professional platforms are baking AI in. You've probably all seen it, regardless what industry you're in, whether you're a lawyer and you're working with LawWare, which is now going off and looking at case histories and all kinds of stuff for you, whether you're in accountancy and you've got Sage doing its AI analysis. I think most companies, they're probably using Xero. That's a really good example of baked in AI. Xero is just full of AI. They've brought out their Jax assistant or Just Ask Xero assistant. They're doing automated bank reconciliations. They do character recognition. They do AI-driven insights. There's a whole load of stuff that's basically been baked into every single platform.
So although we're saying there's all these cool tools out there; you can go get Microsoft Copilot and you can go get build agentic AI that we'll talk about in a minute, you can build agents to do stuff. You'll probably find that you're already using, or have access to, a whole load of AI within these tools. And to be fair,
that's actually one of the best ways to use it, because it will fundamentally speed up the efficiency of what you are doing in your industry-specific business. So it's great having all these tools in Microsoft. It's great being able to go off and ask Copilot to do things and analyse things for you, but if your data is already sitting in a CRM, for example, it's in HubSpot, that AI is designed and built specifically to look at that data which is in that platform and will already be built to give you back useful reports and do all the kind of stuff you might want to do with it. Realistically, that's probably where you're likely to come across AI in the first place. It'll be in your platforms, albeit they probably want to charge more to use them, but they're there.
Stacey Williams
I was just going to add on to that actually, that this is a really good point that kind of comes back to what we said about what AI is. It is just looking for patterns and data to feed you a response that is generally the average of what is likely to be accepted. So when you're asking, you know, a tool for help to build a HubSpot workflow, it's only going to look at the HubSpot FAQs in the first instance. And so
these industry specific softwares already skipped that step. They know where they need to look. They know where the more likely accurate response is going to come from. And so I think that the industry stuff is really interesting for that reason. That was all.
So we wanted to talk a little bit about chat versus agentic AI. We know that chat AI is helping people work faster, get more answers, et cetera, et cetera. But agentic is the next leap in things, if you will. It begins to do parts of the work on your behalf within defined limits, with a caveat. And I'll talk about that caveat in a minute.
What Agentic is doing in comparison to chat AI is instead of responding to prompts, it's going beyond responding and carrying out multi-step tasks. A really good example of a tool that is doing this right now is an AI coding agent like Claude AI, where it's actually developing for you, it's producing code, not just telling you how to do it with a step by step, which is what chat AI would do. So where your chat AI is really good for research and writing, brainstorming, et cetera, Agentic AI is much more useful for that productivity gain. Workflow execution, task coordination, it takes away the operations of things and gives people the means to step in strategically.
You can set it to run, but essentially it still requires that human judgement piece. And that's a really important part of all of this is that at the end of the day, AI has a huge place in business and life alike to add to productivity, but still requires that human in the loop piece. And so the key leadership takeaway here is that chat AI is great for improving individual productivity, but agentic AI is actually helping us redesign how work gets done. It's speeding up teams, it's speeding up workflows rather than just single tasks, which is really exciting.
And it's important to talk about why this matters before we can even say there's a good business case to do this, because it will vary business to business. And there is a great business case to use AI if you want to improve your business productivity and outputs by 40 something percent. But it comes with many caveats, which I'll touch on in just a second, and also requires an understanding of it relatively deeply, having people in your team that get it. Agentic AI does act beyond chat. It interprets, it breaks tasks down, it executes multi-step processes, as I mentioned, but it needs built. It needs a person in the loop that understands what you're trying to build. And it can offer suggestions on how others might have built it or best practise research and things like that. But essentially, in practice, it will gather the information, update the systems, draught the outputs, etc. It will ask for review. It will probably require you to publish in many cases.
But the leadership priority here is that target bounded high volume tasks with guardrails, human review and clear permissions being super important. And just to come back to that caveat, you know, there's recently been an article floating about a small business founder whose name escapes me right now, but essentially, he built a workflow using Agentic AI, and the AI went and wiped a huge amount of customer backups, and there were many reasons for this, but largely it was he gave the AI very specific commands, and the AI not only ignored those commands but had access to things it shouldn't have. So it can go wrong. It can be really problematic, but I think ultimately it has so much potential that if it's done properly, it can be super advantageous for your business. And we're going to talk about what done properly looks like in just a minute.
Alastair Struthers
So...
Leadership use cases. It's a really interesting one, this one. We've been, I would say, somewhere between evangelical and terrifying on what AI can do. Because ultimately, at the moment, it is both terrifying and exciting. It can do some amazing things, but there's a lot of stuff that can go wrong, particularly if it's not done well. And we all have that fear of what if we don't do it right? That example Stacey just gave of the company that had all the data wiped by and rogue AI, it's not the first time it's happened. It's not going to be the last time it's happened. So there's a lot of thought and planning needs to go into how you connect an AI up and what you do with it. But there's some early wins that can be had with AI implementation. And really for business leaders, this is the kind of stuff you want to know about. The really simple things are things like reporting. You know, AI is really good at looking through all this data and pulling together reports. We have recently put together an AI agent that goes off, looks at all of our numbers, and tells us where we're performing well, where we're not performing well, where there's a weird abnormality in our accounts, we need to go and have a look and see why something's been coded in the wrong place. It just has that oversight of the business that we might not get by reading six or seven different reports. It's really good at reading 6, 7, 8, 9, 10 different reports, or even better, plugging into the systems that generate the reports in the first place, reviewing that and amalgamating all that data.
It's fantastic. We actually have a customer, they build hotels. They've got reams and reams of heating information, just statistics about how hot all the rooms are in all these hotels. And it used to be somebody's job to spend, well, actually he's one of the really high level people in the company, I should say, he used to spend about a week going through all of these reports and manually trying to collate stuff and figure out where they could be making savings, where there were leaks, where they needed insulation, all that kind of stuff.
He now takes all of these sheets, chucks them into Copilot because they've got a fully licenced Copilot, and literally in about 5 minutes, he gets back this result of what used to take him a week to do. I mean, there are huge wins to be had, and that's just one example. There's a lot of creativity that can be leveraged with AI to massively increase reporting, help you make better business decisions, and then ultimately, hopefully helps you be more successful. But you can also use it for marketing. You know, one of our peers, I spoke to him the other day, he was saying he's using Claude to do a lot of his marketing. It's writing his social posts. All he has to do is just check them. But not only that, it's also doing his pay-per-click advertising.
It's reviewing and seeing what's working, what's not. It's then making suggestions to him on what to change, how to improve that pay-per-click. He's gone from spending, I think he said £1,500 a month on pay-per-click to £40, and he's getting more referrals. So again, there's huge wins to be made just by leveraging the power of AI. And then you've got that whole workflow piece where you can automate a whole lot of process from start to finish. We at the beginning of that journey ourselves - of automating a lot of things. When customers put requests through our portal for us to do things, we're now looking at “can we get an AI to automatically triage that, can it automatically categorise it, put it to the right place, put it in front of the right person, just to speed up all of these processes”. You know, we're not looking at replacing people here. We're looking at making people more useful. And actually,
now is probably quite a good time to talk about why, as business leaders, we need to be making sure our staff aren't worried about the idea of AI coming in. I mean, I was speaking to one of our other customers this morning who was just saying that their staff are quite reluctant to talk about using AI because they think it's going to come and take their jobs. And I gave her the story of radiologists. When AI first came out, we probably all saw the stories about the fact that it was really, really good at looking at, and interpreting x-rays, and was better than people at spotting abnormalities. And all the radiographers in the country went, oh, we could lose their jobs! But actually, what ended up happening is because the efficiency of them got better and the accuracy got better, the demand for them went up and they've had to hire a whole lot of new radiographers to supplement the demand. You know, AI, it's an enabler. It's not something that people need to be scared of. And I think we need to make sure that the staff and our businesses are aware of the fact that it's there to help. It's not there to replace them because fundamentally, certainly at the moment, and for the foreseeable future, you're not going to be able to replace people with AI. All the companies that made all these huge layoffs a couple of years ago because they thought AI was the future are hiring people back again because they realise you need people in the loop. AI is a tool, it's not a replacement for people.
We talked at the beginning about AI readiness. AI readiness is a huge topic. It's why we've got a whole proper Webinar about it coming up. It'll be the next one we do. But I just wanted to very briefly say here that
It's not something we would recommend you just go and turn on tomorrow because AI is stupid. Fundamentally, it is a toddler and it needs to be taught everything. It needs to be taught the context of everything. It needs to be told what it is allowed to do and what it is not allowed to do. And there's a load of control that needs to go in there. But from a very simple, high level point of view, you know, before you do it, you need to figure out where your data is, who can access it. You need to clean it up and remove all the old stuff that's going to confuse an AI. It could come across a standard operating procedure you got from 2002 stored in a file somewhere, and it doesn't know that that's no longer relevant. It thinks that's current. So there's a whole lot of data cleaning up you need to do. You need to fill in your gaps, and then you need to link your systems, you need to build your controls.
And then there's the governance piece. Because it's big. It's not a simple thing. You know, you need to make sure you've got AI policies in place for your own staff so that they know what they can and can't do, what's permitted, what's not permitted. You need to review your employees' actual contracts and their employment, because that needs to reflect whether you're using AI to as part of your business, whether it's affecting their jobs, whether you know, things like healthcare or payroll’s going through AI, whether they're allowed to use it, what the repercussions are if they do use it incorrectly, the all kind of stuff. You've then got your clients contracts, and you need to tell them whether you're using AI, you need to cover off the whole IP thing about who's generating what and who owns what. And then you've got the whole governance bit of just making sure that everything is aligned and ticked and ready and you've got your risk reviews and you've got all that stuff in the background. And that's before you turn AI on. Because otherwise, you're likely to get yourself in some deep hot water, I suspect.
I think that's pretty much us really at the end of our webinar, which is good. We're just over half an hour, so we're there.
This slide is just a reiteration. There are the dates for our next webinars. We're going to rerun this one just in case anybody missed this one. We're going to do exactly the same content again on the 10th of June. We'll see if we can get some other people along.
But if you want to join again, feel free. If you want to send any of your staff along to rewatch it, feel free. We'll circulate the booking link and then we'll circulate the booking links for the other ones as well and at the same time to everybody who signed up for this one. So we'd love to see you. We can do a deeper dive on actually getting your business ready and how you can use Copilot more efficiently. And then we'll look at rolling it out.
So that's really it for the main content. We've got a Q&A at the end. We literally have a minute, but I'm happy to hang around for as long as you are. So if you've got any questions, feel free to pop them into the Q&A or the chat. Frankly, we can pick them up on either and we will answer anything we can. And if we don't know the answer, we'll find out and come back to you.
Stacey Williams
And if you don't have any questions, thank you so much for coming today. It's been, I feel, a very positive run-through for us. And it's been nice to have an audience that hopefully has found it useful content.