Unlock the power of strategic AI for your business

00:00:02:21 - 00:00:20:24

Speaker 1

Good afternoon, and thanks for, attending our AI masterclass with RMIT Online and Hypergen. Before we jump straight into the content, I'll just provide a welcome to country. So we acknowledge the traditional custodians of the land, the Wurundjeri Woi-wurrung people of the Kulin Nation.

00:00:21:01 - 00:00:43:17

Speaker 1

And we pay respects to the elders past, present and emerging and whose unseated lands RMIT Online and Hypergen conduct their business. So let's get started. We are two minutes in and we've got some still people joining, but it seems that everywhere you turn someone, somewhere is talking about AI and not just in the context of replacing jobs.

00:00:43:19 - 00:01:15:24

Speaker 1

The real shift that we are seeing is at the organizational level, it is reshaping entire workflows, decision making processes and how we think about productivity. And it would be a mistake to think of this as another implementation. It's really a full transformation. And how we think about work, and it's happening fast. In fact, Gartner predicts that by 2026, over 80% of enterprises will have used AI agents or some AI automation tool to redesign at least one core business process.

00:01:16:01 - 00:01:41:19

Speaker 1

Companies like Microsoft, Google and SAP are investing billions. Well, 250 billion, to be precise, to embed agents this year across their systems. Now, today's session is only 30 minutes, so we will keep this quite focused. You'll see some demonstrations in action, explore how to identify some quick business wins and walk away with a simple framework to assess potential AI projects in your own context.

00:01:41:21 - 00:02:12:11

Speaker 1

So let's be clear this is not a session about ChatGPT. We're not diving into prompts and model parameters. Instead, this is going to be about sparking ideas, getting creative, and helping you leave with ideally a clear sense of where AI and agents can actually solve real business problems and why they're different from traditional automation. And future sessions will unpack ethical dilemmas that AI presents, explore what organizational maturity looks like in this space, and talk about the new skills that our workforce will need to.

00:02:12:11 - 00:02:44:10

Speaker 1

remain not just relevant, but competitive. Our speaker today is Alex Papli. Alex is the principal AI consultant at Hypergen where he helps businesses design and implement AI solutions that drive efficiency and reduce cost. Previously, Alex was Microsoft's data and AI lead for several Australian enterprises and with a deep expertise in identifying and prioritizing AI initiatives, he's played a pivotal role in shaping data and AI strategies that present business value.

00:02:44:12 - 00:02:59:10

Speaker 1

Alex has also designed a groundbreaking approach to training vision based AI systems, earning him recognition as the inventor on an associated patent in 2020. So I'll let Alex take over from here. Thank you.

00:02:59:11 - 00:03:20:22

Speaker 2

Thanks. Aiden and thanks for, the opportunity to present and appreciate everyone, making time out of their day, to have a listen. As Aiden said, I want to we want to keep this really focused, and short’n’sharp, and to really give you some tools to inspire you and think about how you can use AI within your organization.

00:03:20:24 - 00:03:46:07

Speaker 2

And so today's agenda, I'm going to just very quickly touch on the opportunity, that AI presents. And what I've seen, in my travels both previously at Microsoft and now with Hypergen then I'm going to jump into the business cases that we're seeing, specifically, the categories in the drivers that are consistently coming up because I think they will, you'll find that they will resonate with some parts of your organization.

00:03:46:09 - 00:04:08:08

Speaker 2

And, you know, organizations you might be working with. From there, we're going to jump into some demos. I'm a really big believer in a picture tells a thousand words. It's not a products, sell. So they're not going to be extensive demos, but they are intended to give you a bit of a penny drop moment, and help you visualize and see how you might use AI within your organization.

00:04:08:12 - 00:04:30:17

Speaker 2

And finally, we've got, at Hypergen, a matrix of 6 different, sort of areas we think about when you assess opportunities for AI. And so we're going to introduce that as well. And then there'll be some time for questions. So keeping that in mind I get asked quite frequently actually. Alex, why did you leave Microsoft?

00:04:30:17 - 00:04:52:11

Speaker 2

Why did you resign from Microsoft and set up Hypergen. And I think sometimes my wife, asks me the same thing, but. Fundamentally, the way I think about it is I do see, generative AI particularly. And that's the AI that powers ChatGPT and other tools like it. It is quite different and quite novel, and it's it's really opening up opportunities for organizations.

00:04:52:13 - 00:05:19:22

Speaker 2

So if you think about it for as long as businesses have been around organizations, we've been trying to think about how can we streamline processes, how can we automate them. But then what happens is that we hit something that slows this down, and it's often going to be having to correct something or edit something in a word file, its's reading information, maybe from a PDF, it might be listening to an audio recording or, you know, looking at images and all those things we couldn't really do anything with previously, right?

00:05:19:22 - 00:05:40:21

Speaker 2

When it came to automation. So had this issue that we wanted to do these you know, improve processes. But then we hit a bit of a stop. And specifically, generative AI, enables us to actually, take this, what we call unstructured content and put some structure and meaning behind it and help re-articulate it in different ways.

00:05:40:23 - 00:05:59:12

Speaker 2

And so what that means is it does prevent, presents a really big opportunity for organizations, because the things that we couldn't do previously now can be done. And I want to be really clear too I'm not trying to overhype it. There is a lot of hype in the industry right now or in AI. So what you’ll see hopefully is very pragmatic approach to it.

00:05:59:12 - 00:06:22:04

Speaker 2

But I think the key thing is that we are able to deal with data that we just previously couldn't work. And so that is quite exciting. So moving into the themes that, that we're seeing, there’s about seven different categories that consistently pop up, labor savings is one that everyone talks about. I'm personally not a I just don't see that many people losing jobs.

00:06:22:04 - 00:06:48:09

Speaker 2

In fact, I haven't seen any off the projects that I've done. So I don't actually see AI is necessarily directly doing that. And a CFO for a super fund that I spoke to recently articulated it really nicely. He sees AI as a way to take away the urgent work so they can focus on the important work. And so the first example, I'd say, you know, from a saving in that perspective is actually one that I've seen with call centers.

00:06:48:11 - 00:07:05:09

Speaker 2

So if you think about it as a consumer, when you call a call center, you know, you might have a conversation for eight minutes, but then there's a piece we call the wrap time, which is the time after the call is done where the person who answered the call has to make notes. Because if you think about it as a customer, I want to make sure that they're writing notes about what we've spoken about.

00:07:05:09 - 00:07:23:08

Speaker 2

So next time I don't need to re-explain myself. And so that's usually about three minutes. And so one of the really nice things we can do with AI is automate that part. And so if you if you're ever calling in a call center, you often hear that message, you know, this call will be recorded for quality and training purposes.

00:07:23:08 - 00:07:43:12

Speaker 2

But what we can do is take that recording, turn it and take a transcription of it into text, and summarize that text and write it to a CRM. And that that in itself, by doing that, we can save off in a couple of minutes. And you know what organization I worked with saved, I think two minutes a call and, you know, just round numbers if you're taking a thousand calls a day.

00:07:43:18 - 00:08:05:09

Speaker 2

So 2 minutes or 2000 minutes adds up pretty quickly. And interestingly for them, they all the dashboards went green when they lit up. And so their customers are pretty happy because they're not having queue times anymore, and they're able to do a lot more. The next one is around third party fees and reducing that. So if you think about it, a lot of times organizations need to outsource different tasks.

00:08:05:11 - 00:08:34:13

Speaker 2

Might be audits, it might be other pieces of work. And we're seeing demand for organizations to say, what can we do to descope the, you know, the consultants that we're working with to just get them to add the value that we need and try to pre-prepare as much information as possible. And so there are business cases I'm seeing across even things like construction and other industries that might be, audit intensive and for, you know, adhering to policies and the like whereby they're really thinking about how can they actually focus in the specialist

00:08:34:15 - 00:08:55:04

Speaker 2

they're engaging on a specialist piece of work rather than, you know, doing everything. And the way they're doing that is thinking about how can they generate much better quality briefs and more information, like more easily accessible. I think one of the overarching motivations, is this third one around business valuation increases. And that's very, much the case in private companies.

00:08:55:04 - 00:09:17:01

Speaker 2

So we're seeing that where companies can be more efficient, do more with what they have, ultimately their multiples go up. And so it's been a very very big motivator for business valuations. The next one around churn, you know, an example I spoke just before around the call center. This is a good one where if you can, you know, effectively, have your customers call up and speak to a human.

00:09:17:06 - 00:09:38:09

Speaker 2

I think in many cases now, they're actually surprised they're not waiting and listening to hold music. So that's quite nice. Additionally, where you might find that your customers depend on you, but then they need to do a whole lot of paperwork we’re seeing organizations that are in more of a commodity space where their customers have to generate all this paperwork off the back of what they do.

00:09:38:11 - 00:10:05:22

Speaker 2

I’ve seen companies exploring the opportunity to actually provide more of that as part of their service and in doing so, keeping their customers happier, but also potentially, adding a bit of a competitive differentiation. So thinking about that value chain, for your customers, and thinking about how they might want to work with you on what they need to do outside of working with you to see where you can use AI to actually improve that customer experience is one that we're seeing quite a bit.

00:10:05:24 - 00:10:27:12

Speaker 2

The next one, production increases, is really the term I like to use is document cracking. And that's basically being able say how do I at scale, you know, read for example PDF documents and understand what might be in them. So we're seeing opportunity within mining and resources whereby they could be a whole lot of different reports that might have been written over a long period of time.

00:10:27:14 - 00:10:50:03

Speaker 2

And they need to extract key data points or measurements out of them so they can actually better articulate I’m sorry, better allocate resources to where, you know, opportunities are. Equally in financial services, if you think about, if you think about investment houses, they might be going along and saying, right, well, we want to analyze all these disparate reports that are out there.

00:10:50:03 - 00:11:13:00

Speaker 2

So industry reports, company reports, even podcasts. So directors are jumping on podcasts. And they inadvertently and fairly innocently say something that might actually be quite helpful, as a, as a guide for an investment decision. You know, we are seeing that there’s quite a lot of companies doing that particular at the bigger end of town, for that. The next one is competitive pressure.

00:11:13:06 - 00:11:38:23

Speaker 2

And that's in some respects, you know, when we think about reduction in third party fees if you're if you're the person who's or the organization that's engaging people. On the other side is if you are an organization as a service provider, and you look at it, you think my customers are probably asking us to go along and do more with AI, or they might try and exhaust this capability or our competitors might beat us to it.

00:11:39:00 - 00:12:01:08

Speaker 2

We're seeing a very big motivation for organizations to actually go along and and actually respond, in a way that can just keep them competitive and, obviously first mover advantage is pretty helpful as well in being able to seize more, customers too. Finally, safety is a really important one, and it's not a dollars and cents ROI.

00:12:01:09 - 00:12:31:03

Speaker 2

but it's a really, really important metric. And so using the vision based AI solutions, where, we’re able to help organizations, you know, do things like improve safety. An example I'd like to refer to is, a company sort of in the logistics. And if you think about it, lots of vehicles, they might be leaving warehouses with, loads on them, and they need to make sure they're all safely fastened, and they're gonna have a process, of course, where someone's taking photos and, you know, they're very clear and saying, we don't want to remove that person.

00:12:31:03 - 00:12:48:04

Speaker 2

It's massively important to us that we we make sure everything's safe before it leaves, or, you know, as go goes off on the road, but what they can do is take that photo and send it to AI for a second point of view. And so they might come back with like a red, green or amber rating on, on, you know, that particular image.

00:12:48:09 - 00:13:11:18

Speaker 2

To then help the person who's making the decision and maybe catch, you know, 1 in 1000 issues that they didn't notice that might just help improve safety. So we're seeing, a lot of motivation for, for, for leveraging AI across those areas. The other thing I'll say, and maybe I should have said earlier, I would always recommend you go and buy before you build.

00:13:11:18 - 00:13:29:09

Speaker 2

And so we would we always say go along. And if there are products out there that do this go and do that. Right. But what we're finding is that's not enough. It's a, it's a rising tide. And if you think about it, every organization, you can go and get a Copilot license or a ChatGPT license and, and get people trained up and using that.

00:13:29:09 - 00:13:51:04

Speaker 2

Well, right. And certainly that's a really important part in skilling up and being able to use those too, tools well, but it's effectively a rising tide. And it’s the new normal, and so what organizations are doing and specifically around these ROI motivations around saying, how do we build our extra IP, extra capability that differentiates our business beyond what we can just buy off the shelf?

00:13:51:06 - 00:14:09:14

Speaker 2

And that's really what we're we're seeing progressive like organizations progressing to say, yes, we want to think about how we can use AI as a competitive differentiator. We want to help us give us a bit of a buffer from, you know, from competitors. We want to just run a bit more efficient and those sorts of things.

00:14:09:16 - 00:14:28:17

Speaker 2

So keeping that in mind, I'm now going to switch to a bunch of demos. The reason for this is just to help, hopefully with a few penny drops, in, in, in the call. So, I won't be, going into detail. They're not meant to be product demos. And selling your product, it's literally just to paint a bit of a picture.

00:14:28:19 - 00:14:47:08

Speaker 2

I will say that, Hypergen is a Microsoft partner. And so the tooling I will be using is, the Microsoft suite of Copilots. And that's purely because that's what we specialize in. And I'm going to provide the best demo to you by using the tools that we know. For companies they do like these because they use they're quite secure.

00:14:47:08 - 00:15:11:10

Speaker 2

They're obviously within your corporate environment, which, you know, is, is a big question that comes up all the time around AI. So the first thing I'm going to do is a really straightforward demo, this is, a policy helper. Right. And so, Hypergen, we're we're a small business. We're growing. I'm very, very, pleased to say we've hired two data scientists to help with the project work, and I'm delighted to have them on board.

00:15:11:10 - 00:15:34:04

Speaker 2

They're both brilliant and super smart, but alongside that comes all the policies that we need to write. And so I went and got a lot of policies, obviously, and made them for Hypergen. But as a business, right, we’re a professional services company primarily, and our customers typically take leave over summer. And so what I needed to do was make sure that our policies, had some sort of reflection around the Christmas shutdown.

00:15:34:10 - 00:15:55:17

Speaker 2

And so this year, what I've done is I've put all the policies in a SharePoint directory. And then from there I've pointed a little Copilot which is like a ChatGPT but just over that directory to help people quickly find answers to questions they might have. So for example, if I'm a consultant and I say, can I work over the Summer

00:15:55:19 - 00:16:11:20

Speaker 2

break? And I'm very happy with, the fact that my typing, I didn't do any typos because there's a live demo, and we'll have our fingers and toes crossed. By the way, for these demos, so please, you know, they're all live. And so what we'll have here is this is going to return information about our policies.

00:16:11:20 - 00:16:28:19

Speaker 2

But the thing is that what we'll see here is it's got information around the Christmas down period. And that's specific to Hypergen. And so that's I'm really pleased that's come through. It means it's working well. But you can also see here the leave policy it refers too, so, this is a very, very simple way of being able to find information.

00:16:28:19 - 00:16:48:05

Speaker 2

So if you're an organization with lots of info like, you know, product information or policies and things that either you or people need to find, or maybe if you've got a call center, you'll need them to answer questions for your customers. It's a really nice, quick, easy way to go along and actually create something like that. That makes it just a bit quicker and easier for people.

00:16:48:07 - 00:17:20:16

Speaker 2

The next thing I've built is an NDA review. So, being in the business that we were in, we often do a lot of NDAs. And so, what I have done here is, we've built a very simple NDA reviewer which goes along and we've got six criteria that we want to look at. And I'll just say please review, and you can see I'll paste this file in and while it's loading, what it's done is it's looking at this particular document inside it to understand if the terms, and what's written in there are appropriate for what we're trying to look for.

00:17:20:22 - 00:17:45:06

Speaker 2

Now, it doesn't replace a lawyer, but it does help us quickly respond. And, you know, as a business, we don't we're not at the scale or, scale yet where we do have lawyers, full time. So being able to do a quick sense checks on documents, is particularly helpful for us. So this is the first, the first demo I just want to take you through around using these tools, just to do things a bit quicker.

00:17:45:06 - 00:18:01:17

Speaker 2

And again, when we think about it, using these tools is a way to speed those processes up where previously we're struggling because, you know, things would slow down because I've got to find a document. I go through all the policies I've got to read over an NDA, these are the sorts of things where we can use AI.

00:18:01:19 - 00:18:24:21

Speaker 2

You know, in a targeted manner to try and speed those up. So the next one I'm going to take you through is workflow. So I'll just make the screen a bit bigger. What I've shown you so far is effectively like a chat interface, right? Where I’ve manually typed in something. But the really great thing, one of the pieces that's probably the most exciting to me, is actually this piece.

00:18:24:21 - 00:18:42:18

Speaker 2

It's not an AI agent, by the way, which is where all the hype is. And I'll show you some a bit later, but it is something that's really, I find really cool. So when you think about a workflow right now, we've got along as most organizations have built little workflows and, you know, where they can, you know, have if/then statements and auto approve and all these things.

00:18:42:18 - 00:19:02:00

Speaker 2

But again, this workflow starts with a file. And so in this particular case, what we can do is we can actually inject a prompt into a workflow. And that's what I'm going to show you here. So if you think about it, rather than having someone actually have to manually type out, the, you know, a prompt to say, you know what?

00:19:02:03 - 00:19:22:17

Speaker 2

Whatever they're asking for, we can actually get the workflow to trigger it. This is a really simple example of working with a financial advisor. Typically if you go to a financial advisor, they'll say, can you please send us an extract from your, your bank account? Because we want to go and classify these transactions to help you understand how you're spending and money and where you might need to be a little bit more effective.

00:19:22:19 - 00:19:44:12

Speaker 2

And so if you have a look here, I've just put an instruction saying, please classify this in the following categories. Living expenses, travel, transport. And then we've got a bunch of transactions. Just paste straight in and let me hit the test button just to simulate what it looks like. But on the right hand side, what will happen is you'll see that it's now going to classify these particular expenses.

00:19:44:14 - 00:20:12:12

Speaker 2

Now in itself, this, in its current state won't be that helpful for a workflow, right, because a workflow can’t look at that and get a very clear yes no answer. But if you go back to the NDA example, it becomes quite interesting. I had an organization with about 200 salespeople, and they had four lawyers and they, they up in B2B and their customers, might often present terms and say, look, you need to use our terms for this particular piece of work.

00:20:12:14 - 00:20:33:20

Speaker 2

And the thing is that some of these were really, really low value, like opportunities for them. But they wanted to serve them, and they were finding that the legal team were getting inundated with requests that were low value, that had to be looked at. And so they've been exploring the idea is saying, well, what if we go along similar to the NDA checker, but instead we have like true false against maybe 4 or 5 criteria.

00:20:33:23 - 00:20:56:15

Speaker 2

So a limitation, a liability clause, an IP clause, a confidentiality clause. And if they are all in there, and maybe if the value of this opportunity is less than, say, $500 or $1000 or whatever makes sense, let's just sort of approve it because the risk with, determined is low. And it will actually help us a lot. So that's a really good example of thinking about how can you actually streamline workflows.

00:20:56:15 - 00:21:07:18

Speaker 2

You must have exceptions for the big deal so that you you make sure that you don't run into trouble. But, these are the areas that we, we find a lot of interest in.

00:21:07:20 - 00:21:18:00

Speaker 2

So I'm now going to jump through to some AI agents demos and I'm just going to have a quick drink.

00:21:18:02 - 00:21:34:15

Speaker 2

So, an AI agent, again, there's a lot of hype about them, but the way I think about them is actually it's a combination of what I've shown you. So the first piece I showed you a policy system which is basically, being at a have a knowledge base attached to an AI model, and they ask questions about it.

00:21:34:18 - 00:21:50:15

Speaker 2

And then the second thing I showed you is how I can use workflow automation to go and do things. Well, an AI agent really is just something that has both those things. And I say it's like CHatGPT but with arms and legs. It can now do things for us and the tool I'm showing you is the Copilot studio.

00:21:50:15 - 00:22:11:10

Speaker 2

As I said, we're Microsoft partner. But the intent here is just to sort of show you how we can use it to, to work over an email box. So the problem state which I'm going to simulate here is, say having a group inbox like contact us or whatever it might be. And in many organizations you've probably got these, and they're really simple questions.

00:22:11:10 - 00:22:27:02

Speaker 2

It might be, can you please change my phone number or my address or, you know, the list goes on. It could be insurance where there wasn't enough money in the bank account. The payments failed and they’re emailing saying actually we’ve now got money in the funds or in our bank. Can you go and re-process it.

00:22:27:08 - 00:22:48:10

Speaker 2

So there’s heaps of examples there. And so what I'm doing is I'm just going to show you how we're going to deal with a request, which could be for a phone number change. Now I'm going to jump into Salesforce. The reasons why I’m doing this is that it's important to sort of recognize that although I'm showing you some Microsoft stuff, it's really about saying we can integrate to most systems where they have what we call an API.

00:22:48:14 - 00:23:06:02

Speaker 2

So Salesforce have got them. So here can see my contact details and my phone number here is triple nine, triple nine. And that's where it ends. And if I come over to my agents, down below, this is the message that's actually in an email which I'll simulate, but it's on the bottom, right. I'll just zoom in.

00:23:06:07 - 00:23:24:17

Speaker 2

You can see I want to change this phone number to triple one, triple three is the ending. So I’m gonna just clear that out just so that you know, that, it's doing something outside of the chat there. But what I've got here on the left is actually description. So these are basically a big prompts for the agent that says what it can do.

00:23:24:17 - 00:23:43:04

Speaker 2

And its role is to go along, read emails and see if it can make updates. It can find information from the CRM. It can also update contact details and I’ve also said there's a safety check in here. If anything else comes through just ignore it. Basically, don't do anything I mentioned before I could attach knowledge.

00:23:43:04 - 00:24:04:05

Speaker 2

This is a really simple example. So I did this. But this is where you would put in things like company policies or any other processes that make sense. And then what I've got here is a set of tools and they say, as I said, the arms and legs. So what we want to be able to do and what this is going to is when I run, the trigger, which is below when the new email arrives in the inbox, what I want it to do was read over that email, find out what it's about.

00:24:04:07 - 00:24:18:06

Speaker 2

And it's, you know, I said, I want to update my phone number. And what it needs to be able to do is update it to update the contact. And actually needs to find the contact first and then it'll update it and it’ll send an email. So what we're going to do is we're going to have our fingers and toes crossed.

00:24:18:08 - 00:24:33:10

Speaker 2

And we've got the the demo in here. And this is the email sent to myself, which is that exact text just earlier in the morning, just to make it a bit easier to to show. But I ‘m going to hit start testing. And what this is doing now is this is actually going to run this AI agent. It runs pretty quickly.

00:24:33:11 - 00:24:52:00

Speaker 2

So, I'll let it run and I'll just take you through each step. But fundamentally what's happened is it's looked at, it's identified, as you can see here, my email address, is Alex at Hypergen. And what it's done now is it's actually logged in to Salesforce. And now it's, you know, zooming along pretty quickly and it's sending the email.

00:24:52:06 - 00:25:08:11

Speaker 2

But if I just step through, you can see it looked up my email address in Salesforce, and it came back with the ID of the record and my surname. And just to show you that's true. You have a look at the last three digits, it's QA3. And in the URL in Salesforce, it's pretty cool because I can actually see that up here.

00:25:08:11 - 00:25:26:08

Speaker 2

It says QA3. So I know it's actually gone into Salesforce and it's done that first step. So then what happens is well it read the email and knew that it had to update that contact. So in which case it said the new phone number needs to be 111, 333, because that's what the email said. And so it goes along and makes that change

00:25:26:08 - 00:25:47:16

Speaker 2

in Salesforce then sends me an email saying, I saw that your existing number was 999, 999. And now I'm making it 111, 333. And just to prove it, what I'm going to into is I'm going to hit refresh in Salesforce. And when we click on details you'll see here that my phone number is changed. So that's a really simple example of how we can think about automating really

00:25:47:18 - 00:26:05:18

Speaker 2

Simple tasks that just take a while, and are a bit onerous. So the last thing I'm going to demo is an audio assistant. And then we're going to go to our matrix, and take you through. I'm going to go pretty quick on this. Well, just because I'm mindful of time. But what we've got here is an AI assistant.

00:26:05:18 - 00:26:25:05

Speaker 2

And the way I like to think about is the voice is just another chat. So we can already talk over a chat, you know, GPT or what you saw in a Copilot. Voice nowadays, is actually just another way of talking because we can just transcribe it. And what I found was there's this great, EV incentives the New South Wales government have done for fleet organizations.

00:26:25:07 - 00:26:40:13

Speaker 2

And so what I'm going to do is I'm loaded that document in and I'm going to have a talk to it. Now that's actually available on our website. It doesn't cost just go to the menu at the top and then you can see it there. But I'm just scanning through and there's a little button down below that says start a call.

00:26:40:13 - 00:26:56:24

Speaker 2

But down below you'll see here this set of instructions. And we can ask it information about the program. We can, you know, tell it about our business and see if we qualify. We can interrupt it. We can do a whole lot of things, and then we can answer back and get an email sent to us with some information.

00:26:57:04 - 00:27:04:10

Speaker 2

So I'm just going to start the call now. This could be simulated and hooked into your phone system. Right. So assume that it's not.

00:27:04:10 - 00:27:15:14

Speaker 3

Hi. This is fleet bots, a demo AI assistant from Hypergen. I am here to answer questions about the new EV fleet incentive program in New South Wales. What's your first name and how can I help?

00:27:16:04 - 00:27:22:11

Speaker 2

Oh, yeah, G’day. It's Alex. Can you tell me what's this program about? I heard about it, but I don't really know.

00:27:22:13 - 00:27:28:05

Speaker 3

Hi, Alex. The Electric Vehicle Fleets incentive is a payment scheme that helps eligible fleet managers access.

00:27:28:05 - 00:27:38:05

Speaker 2

Sorry, I'm pretty busy. Can you just tell me, I'm a baker? I'm based in Sydney. I've got three vehicles. Do we qualify?

00:27:38:07 - 00:27:56:23

Speaker 3

Yes. You may be eligible. The incentive has two funding options. Kickstart funding and competitive bid funding. Based on the information you've provided, the kickstart funding may be suitable for you as it supports organizations that want to pilot EVs before a larger transition. Well, would you like me to provide more details on the.

00:27:57:00 - 00:28:08:13

Speaker 2

I've got a mate who's got a very big company. It might be interesting. They've got like a few hundred vehicles. Is there part of this program to support really big fleets as well?

00:28:08:15 - 00:28:13:01

Speaker 3

Yes. The competitive bid funding option may be suitable for different companies as it supports

00:28:14:00 - 00:28:19:00

Speaker 2

Sorry, my son's just walks in the room. Can you just give me a moment?

00:28:19:02 - 00:28:22:00

Speaker 3

No problem. Alex, I'll wait.

00:28:22:02 - 00:28:28:04

Speaker 2

Oh look, I'm back. Hey, I forgot, where are we at? What do we. What are we talking about?

00:28:28:06 - 00:28:32:01

Speaker 3

We were discussing the two funding options available under the electric vehicles.

00:28:32:01 - 00:28:38:10

Speaker 2

I've gotta run, can you just drop me an email with some information about what was spoken about please?

00:28:38:12 - 00:28:42:16

Speaker 3

Yes, I can do that. Could you please provide me with your email address?

00:28:42:18 - 00:28:48:13

Speaker 2

Oh, yes it's alex@hypergen.com.au

00:28:48:15 - 00:28:53:03

Speaker 3

Okay, just to confirm, your email address is alex@hypergen.com.au

00:28:53:05 - 00:29:13:09

Speaker 2

Yeah that's right. Yeah. Cool. Thanks very much. Really appreciate it. See ya. So that gives you a bit of an idea of what we can do. And, it's it's pretty cool, right? Thanks for the thumbs up. But the reason I want to show you that is that when you think about business cases, if we can deflect calls and make it easy for people to self-serve, it's really good.

00:29:13:09 - 00:29:30:14

Speaker 2

But the thing I really like about it and what really energizes me is these, these don't need to be in English and we've got a really diverse, community, particularly for government, where you might be writing content in one language, but you want to make it easy for people to be able to ask questions in whatever language makes sense.

00:29:30:16 - 00:29:52:02

Speaker 2

This, AI voice bot is phenomenal when it comes to thinking about, how can we actually, you know, be more inclusive for our customers and our stakeholders? So I'm on the second last slide. I'm gonna zoom through because I know Aiden is going to pull the plug on me very shortly. So there's the six criteria, that we talked to another one is pretty straightforward.

00:29:52:04 - 00:30:15:02

Speaker 2

Business impact is it actually useful? Number two, frequency is really important. You know, if we're only getting one email every month to change the mobile phone number, I probably wouldn't be doing it. But if we're having having them very frequently, that's a really good example where we might want to build something. Human-ai collaboration is a term I probably would probably need to come up with a slightly better one, but this is actually more around scope.

00:30:15:04 - 00:30:35:08

Speaker 2

So if we find that the solution we're trying to do is a score of one, out of five rather than five out of five, which all of these are, what we're probably doing is we're solving a problem that's too big. So what I generally find is that if we can de-scope it, what actually happens is that we can make that number from a one to a five

00:30:35:08 - 00:30:55:13

Speaker 2

for the human AI collaboration, but also reduce the costs. And that feeds directly into the build effort, which is the fourth item there. Safety is really important. I think first of all, what I've shown you, for the most part, that's all within a Microsoft secure environment. So security is, if you build, build it and design it, right.

00:30:55:15 - 00:31:21:21

Speaker 2

You can do AI safely. But what I mean about safety here is more around brands. So for example, that example with the, you know, EV fleet incentive, a small organization might say, hey, we're happy to have these voice bots running and doing these things to help our customers. But a really big company might say, look, we're just not totally confident because even though it works, if we have like one error in a thousand calls and they go to, you know, Today Tonight or something, it's going to be a really big issue.

00:31:22:02 - 00:31:43:10

Speaker 2

So we do look at it from a safety perspective, mainly around reputational risk, and certainly regulatory. But I've mainly seen more reputational being the issue. So those are the sorts of things we just look at and go, is it something we feel comfortable with? And you might say and instead well, let's just create information assistant that our call center agents can use to then ask the questions quicker for customers.

00:31:43:10 - 00:32:02:09

Speaker 2

So that's the way and the final one's longevity. It requires a little bit of staring into an oracle but it's really about thinking about where's the direction of this technology going. And is it something we should invest in? Because as I said at the very beginning, we would recommend that instead of your, building from scratch you would buy and then you try and build on the top.

00:32:02:09 - 00:32:18:07

Speaker 2

So what we want to do is try and make sure the opportunities we're focused on are ones that are going to have a bit of a lifespan. The other thing is, from an ROI perspective, I've seen I, I've not yet found a company where we can't find a positive cash flow within 12 months of initiatives like this.

00:32:18:07 - 00:32:36:07

Speaker 2

So many out there when you start looking. But what we do generally say it's look, from a longevity perspective, you can think about it as a no regrets investment. So even if in two years time and some amazing tech out there that does the job so much better. If the payback was within 12 months and you got it up nice and quickly, then you're actually ahead anyway.

00:32:36:13 - 00:32:57:06

Speaker 2

So longevity is really helpful lens. Say is this something that, given the speed and pace of our organization and how quickly we can adapt, is this going to fit that criteria? And again, it's a 1 to 5 score. So we can offer some consulting to help you shape these opportunities up and identify them. But that's that's a story for another time.

00:32:57:08 - 00:33:30:07

Speaker 2

But hopefully this will give you a bit of a start. Some of them are really common sense. Some can benefit from a bit of domain expertise like that build effort and longevity. But they should absolutely get you started. So the final slide often companies say, wow, we're so excited we just don't know where to start. And so what we recommend is you do a beachhead, and that's where I break the rules sometimes on the 12 months ROI or the payback just to say, just do something, just get something small, something that's low risk, something that's just going to say as a business we're using AI, we're just doing it for something.

00:33:30:13 - 00:33:46:21

Speaker 2

It could be just even looking at an email, sending it to the right team, like doing a triage, not even updating a system, but as soon as you get something, it's like that realization in the business saying, hey, it's we're using it, we're using it well, we're doing it beyond the product. And we're doing something that's for us.

00:33:46:23 - 00:34:10:16

Speaker 2

And once that happens, you then really open the floodgates to what we call boost, which is really working through that assessment matrix being very programmatic over, let's prioritize dividing the issues. And then even though there's a lot of talk about transformation with AI, I think that if we go down that path from the beginning, it's, it's something that can be a lot of time wasted. What I actually find is some of the initiatives to the to boost can become breakthroughs.

00:34:10:16 - 00:34:29:03

Speaker 2

And one example, I’ve had is we're helping a company right now be defensive in terms of that competitive, labor that I said previously, but by optimizing the way that they actually do delivery so that they can make sure that they remain relevant with their customers, what they're actually recognized since they could potentially sell that as a new offering.

00:34:29:03 - 00:34:48:00

Speaker 2

You need a much bigger market. And so that's certainly falling into the breakthrough space. So anyway, we've rushed through it. I hope this has been valuable. I can see we still got lots of participants online, which is great. So thank you for sticking with us. If you want more information, we tend to post a lot of Hypergen

00:34:48:02 - 00:34:54:18

Speaker 2

LinkedIn, you're welcome to contact us. And, and thanks again. And I'll pass you Aiden for questions.

00:34:54:20 - 00:35:11:06

Speaker 1

Great, thanks Alex. I think I speak for everyone looking at the chat when looking at that conversational agent. It's phenomenal. And the applications are going to be incredible. We've had some questions come up through the chat. Feel free if you are still here for next ten minutes, to place them in there and we'll try and get back to you as soon as we can.

00:35:11:08 - 00:35:23:12

Speaker 1

Alex, one of the first questions not unexpected that we're going to get is around data security, privacy, utilizing these types of technologies, but utilizing it effectively. Do you have a summary for everyone who has those questions?

00:35:23:14 - 00:35:47:16

Speaker 2

Yeah. Look, absolutely. It's a really good question. I think there's been a lot of fear in the market and potentially and rightly so. I think, you need to first of all, be really careful about what tools you use. And so, looking at terms and conditions is important. One of the reasons why, I've sort of focused in on Microsoft, apart from being ex Microsoft, is that we can, by dealing with all the things within the, ecosystem.

00:35:47:18 - 00:36:08:14

Speaker 2

It's a lot easier to deal with security. And when you think about AI, particularly in the world of agents, it's actually talking to other applications. So you actually want to have everything behind the firewall. So, I think it, it is a very important approach. And so what we find is that there are some really great demos out there of things that just look amazing, right?

00:36:08:16 - 00:36:25:18

Speaker 2

What I would say is use those as inspiration because they're usually fantastic. I've seen some great ones, out there, and I'm sure many of you have but use them as inspiration. And go to your I.T team go that you know, people like us if you need to go and just validate. How can we do this in a secure way?

00:36:25:18 - 00:36:46:00

Speaker 2

Because you absolutely can take a copy, even say, you know, OpenAI's models, like they have a thing called GPT.40, for example, and Microsoft get a copy of that. So they don’t actually call OpenAI, they've got a copy of that model and it's inside Azure. And you can call that in the same with Amazon. You know you can do that with AWS and bedrock and a whole host of other open source models.

00:36:46:00 - 00:36:55:20

Speaker 2

So it's really just thinking about security as part of the solution design. But you can do it. You just gotta, you just got to make sure you're talking to the right people.

00:36:55:22 - 00:37:03:09

Speaker 1

Thanks, Alex. We've got one other question here, which I think is quite simple. Can that conversational agent make outbound calls?

00:37:03:11 - 00:37:26:07

Speaker 2

Yeah, they can, they, I think, funnily enough, that's that's a really good one around ethics Aiden. Around how do we think about when we use them, and when we don’t. Yeah, you can program them to do that. I've seen scenarios where, I mean, I look at it and think, if I had a flight cancellation, I'd be probably delighted to have an AI agent called me up and say, Alex, we know your plane’s delayed.

00:37:26:09 - 00:37:46:06

Speaker 2

We're here to help you and figure it out and have a conversation. I'd love that, because those sorts of those scenarios, that's where the call centers, spike and I just can't deal with the volume. So I think, yes, you can do outbound agents. So there's actually plenty of demos of people saying they've built some sort of bot that can do cold call prospecting and things like that.

00:37:46:08 - 00:37:51:09

Speaker 2

Again, just look at your brand and, and how you think about using them. But yeah, it's possible.

00:37:51:11 - 00:38:06:12

Speaker 1

Absolutely. There's, there's always going to be the customer at the center of this. So how would they feel getting it outbound call from an AI agent should always be front of mind. Exactly. We've got so many questions coming through. I'm gonna try and get through all of them. So if we don't answer your apologies, we will try and get back to them as soon as you can.

00:38:06:14 - 00:38:26:17

Speaker 1

I'm going to combine two here. That said, when you start developing an AI governance framework, like where do you start? How integral is it to hit the guardrails, safety pillars. And then in comparison to that, can Hypergen agents you demoed be onboarded into enterprise pilot systems? Do they overlap in any way?

00:38:26:19 - 00:38:49:09

Speaker 2

Yeah, that's a big question and it's a good one. So it's interesting like I’ve found, and this is this is a personal observation. I find that a lot of companies feel compelled to write really extensive governance. And, you know, AI policies and all those things before they actually start using it, and even just doing like that beachhead use case.

00:38:49:09 - 00:39:13:13

Speaker 2

And I find that what happens is they spend a lot of time doing it, and then they need to go and review them or realize they were either heavy handed all that duplicated stuff. So I, I'm very pragmatic in saying that. And like if a customer said to us, you know, we want help with the policy. I'd help them, but I'd actually probably say, let's learn by experience and do something which is low risk and get confidence so we can get a sense of where we need to look.

00:39:13:15 - 00:39:32:06

Speaker 2

And an organization I dealt with, put it beautifully, and this is probably nearly two years ago. They I said to them, you know, how are you thinking about safety and risk? He said, well, Alex, like, you know, our business is safety, like, and they're they're an industrial, organization. He said, like, we've got a safety policy.

00:39:32:06 - 00:39:50:06

Speaker 2

So what we're gonna do is we're just going to add a few paragraphs in that talk about AI safety as another thing, because fundamentally it's part of that. And so I think that there's a temptation to say we need to create new policies for everything. And governance frameworks, when in actual fact, you know, you might already have them.

00:39:50:06 - 00:40:10:22

Speaker 2

And it's just about saying, well, AI it's just another tool. It's just a tool in the toolbox that we've got. But, but it doesn't necessarily mean we need to have completely new everything for it. But notwithstanding that, I think the logical things, governance, you put in controls like, let's make sure that we don't just open it up so anyone can integrate with any system.

00:40:10:22 - 00:40:34:05

Speaker 2

But again, most organizations have already got this. So we have agents go rogue. You can publish, certain components of them and share them around so people can build agents and have reusable pieces that can help promote, good quality governance. But yeah, it'll depend on every scenario, on how we do that. And I don't want to sound like I'm taking it lightly.

00:40:34:07 - 00:40:44:19

Speaker 2

It's very important, but I think it's about sort of working through it and not feeling you need to get to perfect before you even start. And that's maybe my message there.

00:40:44:21 - 00:41:10:02

Speaker 1

Great, appreciate that, it's always going to be key when we look at getting buy in from everyone and making sure that we adhere to the ethical principles of the business. Looking at the last couple of questions that we get, I'm going to combine two more. So the Hypergen license based according to the number of users. How does it work for you and the agents that you demoed, were they using generic data, or were they trained specifically on your internal documents?

00:41:10:04 - 00:41:29:09

Speaker 2

Yeah. So. So Hypergen is not we don't sell applications. Just to be clear, we don't offer a license. We're just we're a professional services company. We're experts in AI. So, we're here to help you build, and implement AI. So, and on whatever tooling you happen to use, as I said, we're, we're specialists in Microsoft.

00:41:29:09 - 00:41:47:13

Speaker 2

So that's our our go to. So it's from a licensing perspective, there's a whole like there's a whole conversation around how you license in the world at Microsoft. That I won't go into just because of time. But now let's just say, the things that I showed you today, with the exception of one of them, I'll tell you about that

00:41:47:13 - 00:42:09:03

Speaker 2

in a minute were all on Microsoft in the copilot space, there's a thing called Copilot studio, which is different to Copilot, that you know in Word. The one exception was actually the voice bot. I actually demonstrated, a product called Elevenlabs. And the reason I think that is, it's just a very, very quick way of being able to show how you can use these things.

00:42:09:08 - 00:42:32:20

Speaker 2

Elevenlabs, generally regarded as the experts when it comes to voice AI. That said, there are issues potentially with and maybe or maybe not issues, depending on your organization. But right now they don't offer hosting in Australia. So, we do need to sometimes look at other models. Again, there are plenty of other speech models out there, that you can use.

00:42:32:22 - 00:42:49:12

Speaker 2

And again, we would look at your environment and say, what's the right one for you? Or obviously we can use Elevenlabs, that's great because it integrates with products like Twilio and a whole host of others, and what we call its IP tracking or SIP tracking. If you're in the world of call centers, you know, you can use that to connect up to that as well.

00:42:49:12 - 00:43:07:21

Speaker 2

So Elevenlabs is the one that was effectively public internet. The other things I showed you, were all private, the final bit about training, funnily enough, when we don’t train models, in generative AI, they're already trained. OpenAI or someone else spent billions of dollars making them. All we're doing is effectively writing prompts and then securing them.

00:43:07:21 - 00:43:29:22

Speaker 2

So even though I, I'm probably downplaying, what we do, it's actually about feeding the right information in and in really importantly, getting search right. So when you think about a policy search, we want to make sure it's surfacing the right things. It's not bringing draft policies and it's bringing in the right stuff. So we specialize actually in search and you know, really helping you deploy AI, securely.

00:43:29:22 - 00:43:31:24

Speaker 2

So that's what we do.

00:43:32:01 - 00:43:57:04

Speaker 1

Thank you. I think that is time. Apologies if we didn't get to your question. We will do our best to try and go through the chat afterwards. But we welcome you to try and connect with us after. We'll try and get back to you on that as well. Alex, thank you for the incredibly insightful presentation. I think I speak for everyone when I say that was super informative and we're definitely going to walk away with a few ideas and what we can use for our own business in our context.

00:43:57:06 - 00:44:06:11

Speaker 1

We'd love your feedback as well. If there are other aspects of AI that you would like to see, let us know. And we can run future projects and recordings as well. Thank you everyone for joining. Have a great day!