Ben Love [00:00:00]:

Everybody, and welcome to today’s webinar. We are talking all things AI today. My name is Ben Love. I’m the founder and managing director of Grassroots it, and I’m very pleased and excited today to be introducing our expert speaker, Andrew Herbert. Andrew is a leading expert in the field of AI into integration for business processes, bringing over 15 years of experience in managing intricate projects across a range of industries, including defense, aerospace, prop tech, and of course, AI technologies. Currently at Remap AI, Andrew has been instrumental in leveraging AI to streamline business processes, positioning the company as a leader in practical AI enhanced solutions. I will hand over to Andrew in just a second. In the meantime, could I please welcome everybody and encourage you to share any questions or comments you have in the chat window for this morning’s webinar.

Ben Love [00:01:04]:

Anything that we are able to answer on the way through, we will certainly touch on. Otherwise, we will return for Q A at the end of Andrew’s presentation. So without further ado, good morning Andrew.

Andrew Herbert [00:01:17]:

Good morning, Ben. Thank you very much for the very delightful introduction and I’m really excited to be here myself. Point to note, going into this, my background from presentation is from military environment, so officers aren’t usually used to waiting for responses. So if you do have any queries or questions, please post them in the chat. I’ll address them on the fly if I’m able, or we can, as Ben mentioned, pick that up at the very end. So today, looking to unlock AI for business, we’ll be touching on session one, which is the first component here. So the content bringing everyone up to speed. I’m not quite sure exactly where everyone sits, but I want to make sure that going into the second webinar, that we’re all using the same kind of language set and we’re all using the same terminology, so we can get a little bit more value out of that.

Andrew Herbert [00:02:09]:

So I’ll be covering off today on what is artificial intelligence. How do we get here? Where is here at the moment with the emerging technologies and AI, then looking at risk and governance and the future, the next two to five years on what we’re going to see on the horizon. So moving straight on into it. And look, from my perspective, I’m always talking about the content. I’m not going to read the content. If you do want to go over this in a little bit more detail, please, the chat slide deck will be available afterwards. Or rewatch the webinar. So fundamentally, artificial intelligence is an umbrella term that covers, as you can see, an array of technologies and a variety of ways that researchers are playing and experimenting with technology and connecting them to each other.

Andrew Herbert [00:03:00]:

The point to note for us today, and probably what you’re going to see a lot in your business environment, is the deep learning and machine learning that top line there, including predictive analysis. And the reason I say that is because machine learning and deep learning, out of that grew the ais, the generative ais, the chat gpts, the bards, the llamas. These models are sitting in this space. And so from a terminology or concept perspective, machine learning is a subset or an area of AI that focuses on enabling machines to learn autonomously. And deep learning of that is quite simply pattern recognition beyond the way that we could even comprehend. So they’re roughly the two concepts in the part of AI that’s a key point for us to remember moving forward, rolling into some key AI terms. Now, I’m going to make an assumption here that we’ve all lived in a house or know of houses, and we can draw this as an analogy pretty well into the AI space. So looking at this, you’ll be able to go over this in detail further on, sorry, later at your own leisure.

Andrew Herbert [00:04:11]:

But remember that the algorithm is the foundational component of the AI. Same way that you build a house, you need to lay the algorithm first, lay the foundation first, and then from there you create that framework, that shell around the model, the house itself, and that sets the use on how you can live and work inside that model and what you can use it for, the training data. And you’ll hear this a lot with training as well, thrown around in terminologies. So training is the actual building of the home or building of the model, and the training data is the raw materials. And so what you’ll actually get is you might get the same house design, the model, the blueprints, using different materials, be it brick or be it wood cladding, sheet metal, a variety of different materials. And when you train and bring it all together, you’ll get a slightly different take on what that construction should theoretically look like. So you can get a same model using slightly different training data to create different quality outputs and different quality generative AI or different quality interactions. Fine tuning is a term that we hear thrown around quite a lot as well.

Andrew Herbert [00:05:29]:

Think of that as renovations. You’re just niching it into something customizing a little bit more than you would expect. Embedding data. Look, I’m biased. I think that embedding is one of the fastest and easiest ways to create value inside a business. In particular, the embedding data is a dynamic aspect. It’s like furniture. You can move it around and change a dining room into a living room, into a bedroom, and as quickly as your wife or your partner asking you to push a couch into the other room, it can be done the same with embedding, whereas rebuilding or retraining a model, essentially new construction of a home, expensive, time consuming, and you may not necessarily be able to undo the work.

Andrew Herbert [00:06:22]:

And also you have to potentially start at the very beginning again. Now, looking at these terms to try and comprehend the scale of what we’re dealing with here. GPT, which is the Chat GPT you’ll hear about the 3.5 turbo from OpenAI and the four. These are all generative, pretrained transformers. It’s a fancy term for saying it’s a really smart home. It’s a really broad model, language model, I’ve got to say. It’s not a math model, it’s a language model that can go out and create outputs. They’re very flexible, broad range of data.

Andrew Herbert [00:06:57]:

And then you hear as well this LLM term thrown around the large language model that is, in a sense, a massive skyscraper, state of the art, but is absolutely huge, and towers above these other smart homes and these other models that are arrayed inside this city that is encapsulated by artificial intelligence. So if you think of Brisbane, although realistically we’re most likely bigger from an artificial perspective of the city, and we’ve got all these houses, these models speckled throughout the suburbs and all these skyscrapers as well, you can kind of start gathering grasp of how big this is within the research perspective and the scope and the potential of what AI is. Again, any questions, please throw them in the chat. I know that I’m punching through a lot of data here and a lot of information. If anyone wants me to rehash or anything, we can catch it up at the end. More than happy to, as Ben said, slide deck afterwards. This is a brief history. I’m really not going to go into the bottom of this.

Andrew Herbert [00:07:59]:

There is reference addresses on the bottom left hand corner of every single slide, so you can read more about it. But really what I’m putting this up here for is to get you to understand or to communicate that this isn’t new. This has been happening for the last 70 years. Artificial intelligence research and development has been occurring for quite a long time now. It’s really only started to hit mainstream in about 2010 2011, when we started seeing Siri Alexa and conversations about autonomous cars. Autopilot from Tesla. All of these concepts are kind of encapsulating artificial intelligence. It’s taking multiple aspects of that initial chart that I showed up, multiple aspects of that, and putting that into a single module, like with Alexa.

Andrew Herbert [00:08:47]:

You can see here now we’re starting to get generative AI, and this is where we’re starting to get some real impact AI and emerging technologies. So this is the 2023 Gartner trend radar. This was actually published in January. And I know this is out of date, but I’ve put this up deliberately, because you can see here that artificial intelligence is a catalyst. It is a bridge across multiple technologies. And there is so much rolling through. The speed at which this develops and the speed at which we gain momentum in a number of these areas is so rapid that Gartner had generative AI. And if you look on here, it’s in the bottom right quadrant at the three to six year window.

Andrew Herbert [00:09:34]:

And this is published in January and here we are in September. And generative AI really started to make an impact around about June of this year. So even Gartner, who is paid consultants to specifically go out and watch this space, missed the impact of the speed at which generative AI would come into our lives and disrupt our business operations. And to build on this as well, McKinsey, who is another consulting firm, they do an annual wrap up around about July each year. And their trends documentation from last year, they had 14 of those trends published in this year’s report. And the only one that was missing from last year, that present in this year, is generative AI. So this has come out of left field. A lot of the specialists, a lot of the tech industry missed the wave that this would come at and the impact that this is going to have on business operations over the next twelve months to five years.

Andrew Herbert [00:10:33]:

Now, as a part of that impact and a part of implementing or even looking at generative AI or AI in business overall, you really do have to look at the risk and your governance strategies. So from a risk perspective, first thing, got to say it, use of AI compliant with legislation. There’s a number of industries out there that have certain storage requirements. I’ve put in their policies around privacy, security and sovereignty. Use of data, storage of data. You’ve got to comply with legislation. First and foremost, there’s a rewrite of the privacy act occurring as a result of the Medibank breach and hack that occurred recently. So watch that space for some tightening up of what you can do with client and customer and confidential data and the variety or the different segments of data, whether you’re holding government or health or financial information, there’s definitely going to be some changes that will be implementing as a result of that.

Andrew Herbert [00:11:30]:

The emerging cybersecurity threats. Just put it on your radar that AI can be used for good and for evil fraud. GPT worm GPT Google them. Go have a look, read up on them. There’s some pretty scary stuff that have been developed on the dark web and the cybersecurity environment has shifted from being this movie themed concept of a hacker sitting in a hoodie on their own in their mum’s basement to being a business model. And so operating a business cybersecurity is becoming more and more prevalent in today’s space of digital assets and digital operations. The third one there, poorly planned integrations causing disruption integrating any new technology into a business operations without any form of longevity planning or integration planning can definitely be disruptive and will generally see a stalling or a stagnation of the adoption of that. And included inside of that is education and stakeholder engagement.

Andrew Herbert [00:12:31]:

Whether that’s change management perspective amongst your staff or however that looks, it’s definitely crucial to go out and plan implementation of any AI generative AI technology into business from a governance perspective. CSIRO noted that only 52% of australian businesses have staff policies. So definitely something for you to go and look at before you go out and put AI into your business or allow staff to use it in one way, shape or form and looking forward. So how big a deal is this? How much should we care? So these are some pretty crazy stats. That is a global value, $4.4 trillion of generative AI and the impact that it’s going to have over the next seven years. Australia alone, 115,000,000,001 of the stats in there as well. The 12 hours, 30% of time saving across AI initiatives that’s been seen over in the US as well. The McKinsey report around emerging technologies replicated a very similar figure of around about 30% of time saving in roles.

Andrew Herbert [00:13:41]:

So this is a major, major disruptor and a major aspect of how we’re going to do business and how business operates moving forward. In the bottom right hand corner there, 80% of jobs can incorporate generative AI in one way, shape or form in order to create augmentation of their tasks to improve their quality output and so allow staff members to really work on higher impact activities or automating those tasks and removing that task altogether from that staff member so that they can then focus on other areas of the business. So look, I really tried to make that as broad as possible and there’s a lot of information there and from my perspective I could talk all day, but I’m very conscious that the real value out of these conversations and these webinars is from you. So if there are any queries and questions, anything from the audience, I’m more than happy to help out in any way, shape or form.

Ben Love [00:14:46]:

Thanks, Andrew. Just while we’re waiting for anyone to throw questions into the chat box, I think the big thing for me is that if it feels like AI is happening quickly, if it feels like it is happening very, very fast, and it’s taking us all a little bit surprised, a little bit by surprise. Look, I think some of the numbers that Andrew shared with us there confirm that it actually is the fact that the generative AI was not on the radar of Gartner even at the beginning of 2023. Or it wasn’t the radar, but it was right out there on the outside and that it’s really taken those expert research firms by surprise as well. I think that’s quite amazing. So it’s probably understandable if the rest of us are still playing a little bit of.

Andrew Herbert [00:15:37]:

Absolutely.

Ben Love [00:15:42]:

Andrew, one question. I know we’re really setting the groundwork and a lot of the foundations here in this particular webinar. This is the first in our two part series and in our second webinar we’re going to be talking a bit more about, I guess, where the rubber hits the road. We’re going to be talking some more about some real world uses of AI that are very accessible and here and present for us all today to start using in our business. So I have a question here, and I’m not quite sure where it fits into this webinar or maybe the next one, but an area that fascinates me, actually is AI, particularly generative AI in the world of education. So it could be around high school, university, any of those levels there. And I am fascinated, watching sort of from a bit of a distance, I guess, but watching to see how the educational institutions around the world are reacting to AI. Have you got any observations yourself about that?

Andrew Herbert [00:16:48]:

Yeah, look, I’ve had a couple of conversations in Queensland, centric from the education side of things. So the education department, without putting formal announcement on their behalf, they are looking at how exactly they can either incorporate AI into what they’re doing or how they can create some frameworks around understanding of how it’s being used within both their curriculum generation and also their students because they are acutely aware of the fact that this isn’t going away, this is the new reality. The same way that mobile phones were disruptive to an entire sector and our lifestyle, same way that smart devices were. So they really need to look at how they’re actually applying it. So from a Queensland perspective, it’s an ongoing conversation globally without touching too much on it. We have a client that is running a private school and they’ve engaged us to look at how to provide AI into their learning environment to enable their students to use it. Not necessarily from a generative AI perspective, around giving the responses and using it to write assessments, but more is around a teacher’s aid and a support for students. So if you think about how chat bots are deployed on websites, and then you take that chat bot, which used to be an individual on the other line of a widget, and they’d have to respond, and if they weren’t there, then it’d be an out of office.

Andrew Herbert [00:18:23]:

And then that evolved into A-Q-A model by using some mainstream chats where you preload some responses and questions that the user can gain access to, essentially an FAQ bot, and now you’ve got AI bots where they have access to a broad range of data that if anybody queries it, can query that broad data and then provide a response. So from a teaching perspective, going out and having access to that curriculum and then having the chat bot in the middle, but having some parameters around that to ensure that it doesn’t necessarily give the response, it actually educates the student around what that response looks like. So yeah, pretty exciting stuff and definitely looking at use cases within education moving forward. And I think it’s definitely here to stay and it’s going to be an early adopter side of things that will take advantage of that.

Ben Love [00:19:18]:

I think there are some really important messages in all of that for us in business as well, which is that AI is here, right? AI is here to stay. And irrespective of what your personal position is on it, if you’re not looking at accepting AI into your business and looking at how you can best leverage AI within your business, rest assured that your competitors are. So that may be very simple things such as the emails that they are writing to their clients are just a little bit more polished or a little bit more rounded than somebody else’s are because they’re using a nice, easy little AI tool like a Chat GPT or a Jasper to help polish those emails before they get sent out, it can be in really simple ways like that. But rest assured, your competition is looking at how to use AI and in most cases will already be using AI, at least in some early experimental form. So now is certainly the time to get involved. And that’s actually what we’re going to be talking about really in the second webinar in this two part series.

Andrew Herbert [00:20:34]:

And look, just in relation to that, there are some mainstream providers of AI that are starting to really mature. Their product base, Microsoft, Amazon, Google, they’re rolling out some amazing tools. They haven’t necessarily had the early release because they’ve been working on it and polishing it. And they don’t release something like a startup does, which is a beta to gain traction because Microsoft understands that their corporate and their enterprise users expect a certain degree of quality. And so it’s taken a little bit of time for them to polish things up and make sure it works. But rest assured that AI is definitely being looking at copilot and others being incorporated into the azure. So the Microsoft universe, and accordingly, you may not realize it, but you’re actually potentially already using AI. And one of the other elements as well is to look at how you can actually use AI.

Andrew Herbert [00:21:32]:

And the big thing would be that your implementation strategy, instead of starting to try and do a major shift like a teacher’s aid, look at minor ways like you just mentioned, Ben, like going and using a GPT to tidy up an email a little bit, make it a little bit more polished. You can give GBTs your preferences, your context, how you speak and talk, and so that it can align those emails a little bit faster and a little bit smoother for you.

Ben Love [00:22:05]:

So Andrew, with that, before we wrap up for the session here, I’d just like to let everybody know on the screen, obviously you can see some information there about Remap AI and of course grassroots it. And those QR codes there will link you straight through to Andrew and myself on LinkedIn. If you would like to stay in touch, please do so and add us there on LinkedIn. Probably the last question I would like to just leave us with, and this is coming in from David in the chat there. What can the average SME manager do next to take advantage of AI? And I’d like to add my own little spin on that, which is what is a good way for the people on this call to start to tune in a little bit more to AI without having to dive deep into the research and the technical body of it themselves. Are there any communities or newsletters or what is the best way for them just to start learning a little bit more gently about the world? Of AI as it relates to them, as it relates to small, medium enterprise and where the opportunities are there.

Andrew Herbert [00:23:20]:

Yeah, absolutely. Look, there are so many resources out there. There are so many places you can subscribe to to get updates daily, weekly, monthly about the changing landscape of AI. I think it can be a bit overwhelming, especially when you’re trying to do your job. So what I’d say is sitting down and having a conversation with somebody who’s across it. It’s a blatant plug for self, but in reality, you can actually, inside your own organization, engage with your staff. Some of the statistics that are rolling around at the moment around staff adoption is like 60% of C suite are already playing with AI and already looking at it and using it. But the only 38%, I think it was McKinsey reporters, actually implemented in their organization.

Andrew Herbert [00:24:06]:

So there’s a disparity there. There’s a gap. So absolute guarantee that people inside your organization are using AI in some way, shape or form. And I’d say engage with them. Adopting and integrating AI, like I said before, the small adjustments, the one percenters, it really is about a compounding effect and the attitude of innovation could be inside your organization to encourage the use of AI and the use of these alternate platforms to see exactly how it is without going too deep into it. I was on a call the other day where somebody shared that they allocated a small budget inside their organization, and they said there is a budget available each month that staff members can pitch to obtain to then trial AI tools, mainstream tools that are available, subscription platforms, whatever that looks like, and they can then go and test them. And the staff members got to write a bit of a pitch to get the funding. And then from there they’ve got to write a review and also a business case on how it would be used inside the organization.

Andrew Herbert [00:25:12]:

And so encouraging those staff to go and actually look at these tools, research these tools, and say, how can I use this inside my job? And then go and bring that back to senior management or to management and say, look, this is what I’ve found. Can we then implement it into a production environment for us to use across the broader business?

Ben Love [00:25:33]:

Excellent. Thank you. So, folks, I hope this has given you a bit of a foundation, some of the groundwork there for AI, where it’s come from and where we see it heading. Please keep an eye out for the invitation to the second in this two part webinar series where we are going to be diving into the real world application of AI into your businesses. What the opportunities are that are available for you here and now today to really start bringing this to bear with some quick wins and also some really cool bigger picture stuff that might get you a bit excited into the medium term. So thank you, Andrew. Appreciate your time and your expertise, as always.

Andrew Herbert [00:26:18]:

Until next time, welcome. Thanks, Ben. Thanks.