Leveraging the power of AI to drive efficiency in your MSP

alrighty uh welcome everyone welcome to this month's msps in conversation uh today we are going to be talking about leveraging the power of AI to drive efficiencies within your MSP so automating and scaling uh within your business and your customers businesses uh so a quick little bit of uh introduction to hosted Network so hosted network is a wholesale cloud and talkative provider for Aussie msps we started as an MSP in 2003 and uh 10 years ago we went 100 channel so celebrating 10 years this year uh we as we started as an MSP we understand why uh it needs to work and we understand the pain when it doesn't so if you're not familiar with MST with msps with hosted Network please reach out uh have a chat but uh today we're not here to talk about us we are here to talk about uh leveraging the power of AI within your business and your customers uh so very quickly though a little bit of housekeeping before I get the panelists to introduce themselves as always this is recorded and will be cut up into both video webinar as well as a podcast so available on YouTube Vimeo Spotify Etc um if you guys have any questions that you'd like answered from our panelists today please feel free to jump right into the chat and we've always loved sort of those questions that come through we've got a friendly bunch of guys here today so yeah without sort of continuing on too much um Matt would you mind introducing yourself yeah sure thanks for having me on um and so I'm Matt Dunn I head up the automation division here at the missing link um where we started off a few years ago by automating our own processes so across our infrastructure and cyber security divisions and then we've expanded this into a client service offering where we either help clients to build their own automation functions or we provide automation as a managed service the automation part of the business cool thank you thank you for joining us uh Andrew welcome yes g'day yeah thank you very much for having me on um so I'm Andrew I'm the CEO of y365 uh we're in Microsoft 365 Focus Consulting for consulting firm and yeah at the moment we are seeing a massive increase in work in that sort of low code no code AI space I suppose particularly in the in that Microsoft space so yeah that's what we do and what we enjoy doing oh thank you and Nick uh the famous author that you are um maybe you introduce yourself thanks Ben and thanks very much for having me my name is Nick beaujad and that is how you pronounce my surname um because most people can't work it out from looking at it I'm the chief software architect at tribe Tech I do a bunch of things I spent quite a bit of my life doing that I'm also a chief software architect for a new automation product called water workflows and the rest of my time I spend my time going around talking to people about AI wonderful well guys thank you for joining us today I'm sure you've got some wonderful light bulb moments for everyone on the on the line um so I guess just to set the tone so we've all seen a lot of press and conversation around AI there's a lot of hype but we're finally starting to see some of the products and startups and Integrations that are coming out of that both msps and our customers uh so you know because we're starting to have these conversations um it's I guess proof that I I guess it's the new Cloud right it's no longer just the new buzzword we're actually starting to see things come out of it so today is really an opportunity for um myself and the audience to pick the brains of our panelists around where they sort of see those those opportunities to um my word of the month at the moment is leverage uh how they can leverage AI uh you know both within their business from a top of the line revenue to bottom bottom of the line um you know profitability uh as well as where they're seeing opportunities how they're tackling it within their businesses and so on so forth so with that in mind let me just stop my screen share and there we go all right so I clicked the right button which is always a good thing so Nick I'm going to start with you because you've got some um some great stories I know for this but uh Andrew and Matt please feel free to jump right in but um let's start with some stories so Nick what are some of the best examples that you've seen and feel free to use a personal one um of AI within your business um you know do you guys run I'll start at the beginning so um you know chat GPT was released to Market really November of last year and um was as as far as everybody says and this is stolen from Microsoft presentation it's been the fastest growing technology known to man um it went from zero to 100 million users in four weeks um which is just faster than anything we've seen before until the next tech turns up and so I was asked to give a presentation about Ai and a CompTIA event and I was I was there in the shower on a Thursday morning going um I can't just get interest ideas yeah definitely the shower so I'm definitely in the shower nothing else I'm in the shower and I'm going oh what should I do okay I know I'll write a book right so the the event was a week away and I went I write a book I'll take a book with me that'll work um so I sat down just got out of the shower sat down in front of chat GPT and said hey chat GPT I'm giving a presentation to CompTIA about chat gbt I want to write a book can you help and it said yes here's the table of contents and I went oh I quite like that but can you write chapter one it went sure and after a lot of control C control V copy paste for those of you who don't know the keyboard shortcuts um I had really disturbings of a book and so about by about lunch time we're ready to publish so I went on to Amazon and the cool thing isn't that that's cool but what's cool is what happened next at Thursday lunchtime I uploaded it to Amazon by five o'clock that afternoon they said your Kindle book's ready and your print books ready and if you actually search my name on Amazon right now and go to chat GPT for executives you can buy this book right now um but that was Thursday five o'clock they said they're ready with the print edition by lunchtime on Sunday I had copies physical copies of a book in my hand and I think if that that's an example of an organization using crazy levels of automation to do something if and I published a book years ago and from type setting and printing and making things happen it was months after you'd release the initial copy and with Amazon if I find something wrong I can go change it online and within 12 hours everything's been updated and people are buying new copies of the book so that's the level of automation some of the big end of town are achieving but for us what happened after that was I've told a bunch of people about it a bunch of people bought it it was a business book it's not hard to hit number one in the best sellers list I think you've got to sell about 900 copies I did 10 times that with this book it was ridiculous it flew off the shells Barnes and Noble now syndicated in the US it's it's crazy but what everybody thinks now is I know loads about AI because I wrote a book on it that's why you hear it exactly so I've been asked to speak here I'm going to Queenstown next week I'm speaking and speaking again in Auckland I'm doing events with hpe and all sorts of people loads are happening for our business it's been brilliant it's been a cracking marketing tool um but that's some of the things you can do with a copy of chat GPT understanding how automation works and making that happen that's a crazy story um and yeah I I was actually with Nick the the Knight that he he wrote that book and so he was you know sharing to come to our event and within the next day it was on Amazon and it was you know I didn't realize that level of automation even existed right when you're thinking it's not just within a process that's to a physical you know environment which is crazy um you know so if that's the upper end of town I'm crazy to see what where we're going to to go to you know that hype and that's what I sort of said at the beginning this this concept of AI and this uh is all this automation level has really leveled up very very quickly um you know and yeah so I guess um Andrew Matt from your side have you what have you seen within as best examples within your customers or whatnot any crazy War Stories you far away yeah no worries I'll jump in here yeah so I mean look we are seeing a lot um I mean it's supposed to set the scene firstly I think um you're absolutely right Nick the chat jbt hype has just been absolutely incredible and and the ramp up to that has been phenomenal um and I think uh you know what we're seeing in the in the sort of business side of things is a lot of businesses now trying to go well okay how can I leverage this technology and get the most out of it to improve the efficiencies or deliver value for my business and and that and the answer that question is different for a lot of different businesses and I think um what we're sort of seeing is a real focus on this language processing component of it but there are actually other facets of of the gbt architecture and and what we're really particularly excited about is the Microsoft co-pilot stuff that does back end off of that gbt um uh large language model but yeah so we're seeing a lot of customers asking the questions we're currently working with a customer at the moment actually in the customer service space uh and for them uh you know customer service queries are you know again bread and butter stuff for large language models because really what we're trying to do is ascertain what information does the customer need and how can we respond to that in sort of a friendly and helpful fashion and develop a little bit of a personality and large language models are absolutely fantastic for that but what it's also really good at is extracting information so this particular customer is dealing with processing of orders um so it's not quite as automated as your book Nick you know in terms of sending out the physical Goods automatically and all of that kind of thing but we're sort of at the point where they've got a uh maybe Legacy is the wrong word like a traditional Erp and stock management system and that it does have a some semblance of an API but by connecting that up to an AI system we're actually able to grab orders um as text you know somebody emails in and says I want three of this thing and five of that thing and um you know and the system can actually parse that and work out what they want and then feed that into the order system and then process through there so that's sort of an example of the type of thing that we are seeing where um really what we're trying to do is is can interconnect systems that we just weren't able to easily connect in the past um and so you know um AI has really allowed us to accelerate that development as well so you talk about leveraging I think we could have built an integration for this particular task um without AI but it just would have taken a lot longer and it would have been a lot more inflexible now with AI we've got the ability to analyze text analyze language a lot more efficiently and quickly and get that integration built really really fast um which is really exciting cool thank you Matt what do you guys what do you think uh yeah so so uh to to next point about where where the story started so I my story started in Microsoft teams actually where so our head of cyber security um posted an article on uh chat GPT and said you have got to check this out so uh started playing with it and then um it turned into training for the the all the company's end users now in terms of how the end users are using it so generally sort of things like writing meeting agendas making Communications more clear and succinct but in it so code development I don't know of a developer who isn't using it as a significantly speeds up the process of writing code and debugging it documentation no one likes doing that so it does a great job of that troubleshooting so you're not needing to go to different message boards to find your answers it does a better job of consolidating what's on the on the internet and bringing that back and also policy review and development so the thing about policies is that they are written by experts and quite often the experts will use jargon and that won't be understandable to an end user for example so making a policy clear and being able to refine your own policy so you can even put information in it and say how could this be better as opposed to rewrite it but then across the different parts of the business so HR writing job ads Performance Management um marketing so they're probably the ones in terms of the uh productivity boost they get the highest level of that and marketing in their keyword ideas content outline generation to the next point about punch out a table of contents and then write the the first draft of each of the chapters similar with content outline generation and I asked um our content developer what in terms of productivity what do you reckon percentage-wise this is this is given you in terms of a boost and she reckoned it was around 300 percent and that was the highest sort of percentage I'd heard of any of the different parts of the business and I asked where do you think that the most time is actually saved you said well interestingly it's actually when you're sitting there with a blank piece of paper and you're needing ideas so you might have some keywords and you're thinking of ideas of what could be a compelling topic and it was more a case of just write me a hundred topics headlines and then look 95 of them might be right but there are going to be five in there that you're like ah no I'm going now I'm into the next step of the process so that's where they were getting the benefits and um yeah and then so similarly to our um uh bottom I remember I said automation we started using it ourselves and then offering it to clients we actually trained clients now in using chat EBT because there are a couple of really important things to know about the the data risks I'm sure we're going to come to risks later but knowing how to use it where your data goes what switches to switch off so that your data doesn't go into the model for example and um yeah generally using it across the business and quite interestingly Andrew mentioned co-parted there so the co-pilot for power automate to develop workflows with just natural language it's Bonkers how much faster that's still in preview but it's Bonkers are how fast that will speed up coding it's it's funny you say the coding because we've got a full Dev team in-house with our devops team and and they use it quite heavily um built into their IDE and those sort of things so it's um the debugging piece of it and I'm a network engineer by trade uh and uh we've actually we've found that the network configuration our CLI component's our scripting as even you know that sort of that bridging that Gap has been very very helpful and we want to sort of deploy something new or it's like hey this is Cisco what is it going to equal in say a 40 gate or whatever it might be and voila there's a config um yeah and and there's a couple of things I think we've just skirted On The Edge on that I wanted um yeah people are listening to understand first of all I've been watching the comments um and everyone's talking about will we have an AI Nick well I don't know if you use the new teams but the new teams has just come out with avatars and those avatars are drivable using an API chat GPT has a whisper API which takes voice and turns it into text you can then use chat GPT functions to say what the Avatar should do and in fact in my meeting this morning they had a virtual Nick which was my voice but the Avatar was acting out what I do for another reason but I'll go beyond that I don't know if any of you have played with what used to be called code interpreter or Advanced data analytics in cat GPT it was funny people say to me a lot what is the first job you think will be got rid of by Ai and I've always said you know glibly um it's not AI that will get rid of your job it's someone else with AI that will get rid of your job if you're not not doing it um and um however or I've been playing with um Advanced data analytics I do a lot of work with accountants and financial professionals and you can take a set of data in CSV or Excel you can throw it at this tool and you can say analyze this and I'll use the story I was at a smbit meeting that they have chapters all around the country if you're not a member you should probably join and and I was there at the organization and there was someone from one of the vendors and I won't say which one he said Nick I've got to write my cuddly Business Report can you my qvr can you help me out and I said well give me the last three year three month sales data he gave me the data I put it into advanced data analytics I went right the qbr for this organization and it wrote the entire report for him with all the customers he should focus on which ones have sold well which Distributors were good which products they should look at and this thing is insanely good so if you haven't you know paid you 20 bucks a month to open AI which is a tax on being efficient and you haven't played with Advanced data analytics you really you have to the thing is super cool I think just to sort of expand on that um Nick that sounds like a really um yeah really fantastic use case for the gbt and I think what it boils down to is um creating a better interface and a better relationship with our systems and our data and I think that's typically been a a point of contention and I suppose a point where a lot of it professionals live their lives in this gray area between the humans and the technology and I think um that is going to be an area where gbt is going to excel because things like um you know data is highly structured very easy to reason about for a computer when we're talking about AI um you know we're talking about programming languages again very well-defined set of rules that they have to follow and um and set of outcomes that the user is after and so bridging that Gap I think is really what has sort of revolutionized it um you know GPT understands what we're saying uh you know and the language that we use and how to translate that into those really structured and regimented formats Andrew got me started right he's an Office 365 consultant and I was just playing with a bit of tech earlier this week um openai have released an open source plugin which is used for vex Vector databases indexing non-structured data which means nothing to anyone because nobody wants to build a vector database but if you've got Office 365 and you're using SharePoint online you actually have one Microsoft search is actually a vector database um and although this is in beta you can plug that into chat GPT now so chat GPT doesn't see your data but it's able to call the vector database and get the right data out of all of your documents and spreadsheets and PDFs and tips and anything you've got that could be indexed in in SharePoint and return that back to you which totally changes that human computer interface for staff in an organization which again is is crazy exciting that's a large deployment where are you saying this along sort of the easier wins guys within businesses um Matt you touching a few around that's an easy Win Ben you know that's that's um anybody who's got their documents in SharePoint true true um okay so let me rephrase the question uh thank you for the challenge I appreciate it um so if we're thinking about uh I'm a tech right and a lot of the guys in the call will have a tech background um you know we've we've used we used sort of the various AIS around our communication improving streamlining making sure we haven't missed things what are some of the approaches that msps can use within their business that you've seen successful around not just sort of documentation those sort of thing is are there is there something that they could walk away today and implement in addition to what Nick is saying so the clarity of communications so being able to put in something there that's going to make your um your Communications more succinct I find that any any email that I have that goes past a paragraph if you put it in there and say make this clearer and more succinct it does a really good job of hitting the mark with your your audience as far as being able to understand the information on particularly in a technical uh role where you're giving technical advice being able to put it and you can say explain this at the level of a high school graduate or you so that it can put it in simpler terms so that the end users get more value from the information that you're giving to them even though it contains the same advice but it's potentially in their language that they can reconcile with I'd say is a quick win yep yeah and I think that's a really really good one sorry Ben uh please just cut me off I love it excellent no I was just going to say I think to expand upon that I think Matt's absolutely correct I think um communication is is sort of a the Golden Goose at the moment I suppose of the GPT capability um but I think something that we have had the capability to do uh for a little while but it's been sort of uh less obvious how to do it I suppose and again you can you can sort of do this with gbt I think is really um managing priorities and actually detecting sentiment so um you know we have been able to pass um you know blocks of text into apis particularly obviously living in the Microsoft World we're dealing with Azure cognitive services but we have had the capability to pass data into that but you know it really it requires an I.T person to build something whereas gbt anybody can ask it hey you know what is the sentiment of this enormous paragraph and so that's really powerful when you're dealing with say helpless requests because you probably want to know pretty early on if somebody's sending a really you know tersely worded help desk request where they're clearly annoyed and you know might need some really uh upfront attention now um you know having those sort of early warning signs or having those triage processes in to go hey this one needs to go straight up to level two because there's something going on there and on that note too being able to determine what is required here is this likely to be a problem that GPT can answer itself is it going to be a problem that needs to go to a a level one set a team or a level two so team or is this this one go straight to the to the uh to the network Engineers so I think being able to extract that those key pieces of information and categorize and triage based on that information is a really quick easy win um that we can get as well is that from an automation perspective or is that sort of a more still more of a swivel chair using the traditional chapters I suppose it depends on on how tech savvy are and how far you want to go with it so I mean you know I suppose I don't want to front run the topic of security and and sort of being aware of that because I imagine that's that's something we have to discuss but I think there's you know that aside um you can absolutely hook up to apis and things like that so if you are reasonably tech savvy um or even using the tooling that that Nick talked about the that's in beta the the uh the workflow generator tools available for power automate and things like that you can actually do most of that reasonably quickly um at least in a test environment yourself but again there's the whole security and privacy and confidentiality discussion as well that is a component of all those type of automations but well let's jump into that in a second but I will sort of throw a challenge to everyone else in the call if anyone knows of any interesting tools or whatnot um or they've you know that they'd like to share with the audience feel free to drop it in the chat I'm sure there's um there's plenty of you know good uh valuable content that you guys are already aware of um but let's jump into that sort of that risk that privacy case but we saw that um you know that the code sample of whoever who was it IBM or someone that's um Sam yeah and you know that's a big risk right what what do we need to consider if we're going to be putting personally identifiable information or rip into these tools I know that there is an Enterprise version coming of chat GPT how can we mitigate that has anyone got any guidance on that okay you'll you have seen that um Italy banned the use of Chachi Beauty as an entire country not that long ago and um that was because they were worried about their citizens not being aware of what information that they were putting into the model so with Chad gbt the the biggest risk is sensitive data going into the model so that Samsung example that you spoke about if there's a switch that open AI introduced in response to Italy's Banning of it because they didn't want the rest of Europe opting out of their new handy tool and there is a switch in your data settings where you can switch that off and and that switches off both your the data going into the model but also your chat history because this the simpler um I guess way of exposing data is that if your user accounts are being hacked and say your users are using their work email address and a weak password then their chat history would be available to the hacker and they could associate that data with your company because they've got the email address so you can mitigate these by going into your settings switch off chat history and opting um opting out of your data going going into the model and the other risk is that chat EBT doesn't know when it's wrong it's simply giving you the most statistically correct answer so it's kind of like a really smart grad that won't tell you that when when they're wrong it doesn't know when it's wrong so it may be inaccurate and the data's only up until the end of 2021 so that's something to bear in mind that the model um is limited and I guess lastly is that it's not only the the good guys who get a boost hackers are using it to write malicious code or craft more convincing phishing emails um so yeah that's I think from my perspective what the biggest risks are I don't know if you've ever heard the term picnic in in the I.T context but it's uh it's very relevant for chat GPT which is problem in chair not in computer between the back of the chair and the keyboard um but very similar so I just expand on that a little bit because there's been a change to the chat GPT terms and conditions on Monday um so one of the things they introduced after the Italy band and Italy hasn't banned it anymore that's been undone um one of the things they introduced after that is there's an additional set of terms and conditions you can sign called the data processing addendum and that actually enforces by contract as well as technologically what happens with your data they also say if you're only using the API so you're not using the chat GPT interface you're using the API none of that data is used to train models or gets stored for any length of time at all which is very similar actually it gets really interesting I'll probably be contentious here when you're talking with Microsoft and they go you should use Azure open AI because it's more secure blah blah blah well there is only apis so are they different I'm not a lawyer I I have no idea but I think the other thing we missed there is as well as hallucinations that getting things wrong and I I describe that like the guy at trivia um who always is convinced he's right for every answer and is wrong and almost all of them chat Chief is good at being that person but it's the bias in the model if you think chat GPT has been trained on what we call the Corpus of human knowledge the Corpus of human knowledge is really biased um if you ask I think they fixed this now but if you you saw us chat GPT to write a story about a cook you get a woman in a kitchen and if you ask it to write a story about a chef you get a man in a professional kitchen that's because of our own biases in our own writing that have been used to train the model as it goes forward so hallucinations bias and then reading the terms and conditions and getting advice as to whether you can do stuff with it or not and we've seen similar things even I mentioned the network engineering example um internally we've actually seen you know just it outrightly getting it wrong but presenting it in a way that is accurate so there's also that that risk as well you know sort of make sure you don't just copy and paste entirely with what you're doing right would be a customer service or whatnot it might accidentally slip in something that's wrong um I think you know um you know Matt you touched earlier on sort of that that training piece of it and I think that enablement is also very important of your team I've actually sent a message internally to go you know what we should probably do some more enablement training to make sure people are using it correctly yeah it's crazy it's the like most widespread productivity tool that has been enrolled out to humanity with zero training so it's the current blam uses for not knowing where their data goes when it's so instantly um useful and easy you know easy for them to put their information and so it's going to be the issue is going to be with your the grad that just joined who didn't know that they were posting your proprietary code in or yeah or like with that uh the lawyer example where they uh in New York um lawyer I got their case written for them by chat GPT and the first two links were to real cases that served as precedence but the last two were completely made up and that lawyer's in a lot of trouble now wow interesting interesting um T's and C's stay you cannot give legal advice and you cannot give Financial advice and then it will say guard rails will come up and say I shouldn't but that's an ethical thing right the thing when I'm talking to people is whatever it creates a human should go over it first before sending it off to someone else that's a really important step in the process um guys shifting hopefully we've covered risk but if anyone's got any other questions around risk and whatnot feel free to to drop that the chat and we can address that um but I'd like to sort of uh shift to we've talked a couple of ideas around sort of Revenue generation and focusing on the opportunity of AI um so sort of expanding on that straight out where do you see the biggest opportunity of AI for msps I think in terms of like we're we're all short on the skills like there's a market uh shortage for skills so I think the the opportunity to be able to do more with what you currently have and also if you're you're people are being driven into the ground working longer hours to try and make up for the the other people that you haven't been able to hire to sort of manage the demand of uh of requests it means that like that work-like balance and maybe even like wild I know but of the four day work week so imagine you uh skilled up in the use of generative Ai and everyone was able to be that much more productive that they could work a four-day week and still get paid the same salary you're going to be a place that hangs on to your people because they're able to be more productive with tools like this and so it's it's the more mundane work that tends to get taken taken care of as a result of it so I think in terms of the the people aspect you might have seen the intergenerational report uh recently about what does Australia look like in 30 years time and we're short on people which means we need to be more productive which seems like and and I was looking at the report thinking oh Shane those poor old people and then all of a sudden I was like wait 30 years time that's me so it's like would you take care of yourself and uh help to make the make the workforce more productive so that we are able to contribute to the economy okay Andrew what do you say is the the I guess the the primary opportunity at the moment for msps yeah I mean look I think Matt's really um outlined a lot of that internal opportunity in terms of really getting more efficient and I think that's that's always going to be the goal of the businesses of a business I suppose is to be as efficient of what we do as we possibly can I think that the AI systems that we've got access to now can certainly help us by um you know responding to to simple queries and providing information um you know much more intelligently triaging and management of requests as I said before I think is a really big area that we can we can really leverage this um but I also think there's sort of opportunities um for msps to become more involved in that process automation space I mean it's been something that that we've really focused on and really learned the tools in the Microsoft space but um you know to a point some of that is going to become a little bit less relevant now because for the simpler tasks the AIS can build those workflows for you they can build those power apps for you and so it sort of starts to put the tools you know democratize I suppose access to the tools and it means you don't have to have a Dev team to to manage that you don't have to have necessarily work phone Specialists doing that stuff um so I think there's a huge amount of opportunities there that are definitely going to eat into into parts of you know where we see ourselves as experts I suppose and I think that is an opportunity as well you know there's an opportunity to just do more uh and build more in that space um without getting bogged down in the code and the Integrations and all the stuff that GPT is really good at now and we don't have to worry about Nick what are you saying so I'll add a couple of things to that I I think I think Andrew's got a really good point um but Fool's rush in a little bit with this for those of us I'm I'm older than all of you right so I'm in that old demographic um but I remember the days of Lotus Notes where anybody and their brother could go and write a mission critical application and I remember moving customers off Lotus Notes where they went ah we've got 9400 Mission critical applications written in notes what do we do with them so we need to be careful we don't get that kind of app spread we need to manage that a little bit but I think the thing is for an MSP and I think anything that's ripe for automation with AI is anytime you're copying something from one window and putting it into another whether that's creating a new user from a help desk ticket creating a billing entry from a PDF from a domain provider all of that can be automated with some automation tools like pair automate or what workflows that we build and it's why we build it um and some AI mix in the middle and if you can do that for your business and I would say it businesses are slightly easier to do it for than other businesses then you've got a real opportunity to offer that to your customers and for msps that's moving you up the value chain away from you know what a lot of people are doing which is router switches desktops laptops and Office 365 into actually going we're not just a cost we're improving our customers businesses and I think that's the really exciting Way Forward and I would say if you're not thinking of doing that then there's probably an awful lot of your competitors that are yeah going deeper with customers right rather than just that sort of the traditional means we've actually had uh fuzzy uh added comment in here which is you know he's usually as a panelist so as an attendee uh it's always interesting him being on the other side but he sort of said uh I am all for AI to be clear but uh we're not called we're not also causing a lot of msps and businesses an issue with uh Shadow I.T I always love that term Shadow I.T sounds so dark and mysterious um but um what do you do you guys say that's causing more issues so the businesses that have just jumped in you know maybe not without you know your guidance jumped right into the deep end have they caused themselves issues that they hadn't thought about for you know you guys as msps I think it's similar as as the cloud um you know we had a lot of customers that got really excited about Cloud products and ended up with their data strewn everywhere um and I think you know we are likely to see some of that shadow iot stuff but what I would sort of say to counter that I suppose from my perspective is that um it just changes the focus for the MSP and where we should really be putting our efforts so as a bit of an example you know when um you know covid hit sort of accelerated that the remote work Trend and all of a sudden we had to change but it wasn't really all of a sudden I suppose but it you know in an accelerated time frame we had to change from focusing on networks and firewalls and servers and all that sort of stuff to focusing on endpoints and it really pushes the protection out to that endpoint device and and management there and I think we're sort of seeing the same thing here it's going to be less about the database uh that it's in it's going to be less about the tool that it's in or the tool that's you know the workflow Integrations and I think we're going to see more of a focus on our data and our information how are we storing each row what permissions are we are we looking at and how we allowed to interact with that the tools that go on top of that um you know I think are all there to enhance productivity as long as the organization and the IT team within that organization or their MSP are really in control of where that data is stored um I think there's a lot of benefit um you know to come out of that I think that's interesting if if like Nick said that you can go deep with the customer whereas if you're just skirting around the edges and you don't have that visibility and all of a sudden you kind of get that oh hey we've done this I think every MSP on here is you know they've got a customer that just randomly signed a contract or something and then it's the msp's problem right oh cool we've got to go do this do you see that as a an issue maybe Matt Nick or Andrew I I think um Ben I think one of the things that my view always has been like this is if you have a shadow I.T problem then as an I.T provider you're not delivering what the business needs so it's not that the business has their priorities wrong I think you have your priorities wrong I'm being a bit contentious there and a bit direct but you do that on webinars right um I think if areas of the business are going out and whacking their credit card on products and solutions technology products and solutions then you need to look at why haven't we why aren't we meeting the needs of the business and you need to get closer to those businesses to understand it Shadow 18 might be an I.T management issue but I think it's an I.T management issue delivered by perceived lack of delivery or performance or provision of services that the business needs so you need to have a really cold hard look rather than just go it's Shadow it uses a badge I don't think that's the right answer I think you just turned Shadow 18 into an opportunity Nick there you go guys there's a pipeline generation tactic right there if you've got a shadow I.T problem it's now an opportunity Matt you look like you're about to say something but no yeah so as far as um I guess you could like Turner also into an opportunity Shadow IIT users who use generative AI could have could could do what they're doing anyway in a more considered way so if they knew about how to ask the question about what are they what are the risks of using this or compare this to other software that you were thinking about and they might make more informed decisions if which if they weren't going to consult I.T anyway at least they're getting um another opinion on what they're doing yeah I I also think there's a communication opportunity here as well so I I think Nick your your statement about um you know it be this being an I.T problem I don't think is is um wrong at all I think that's absolutely correct I also think there's an element of communication here as well though it's not necessarily always that um you know it's not providing the service that that um that that this particular component of the business requires it it might just be that they're not aware that they can provide that service or that the communication breakdown's there and so to flip that down into an opportunity as well um you know we've already talked about the benefits of of uh using AI to help with communication with our customers I think that's a really massive opportunity to make sure we are keeping consistent regular high quality uh communication channels open with our customers and providing them with the really critical information they need to help mitigate some of that shadow I.T problem yeah for those of you that do qbrs maybe even adding in you know sort of understanding that shadow I.T you know but also that this opportunity side of it as well going deeper asking those questions I like to say to my team I'm you know that host network with tongue-in-cheek but ask the question right you know actually get in front of them and go hey what are you doing with this why are you doing this okay cool all right don't don't attack it but actually just challenge it go deeper so that you can actually understand that and I think that's that's the biggest opportunity it's always easier to you know drive more revenue from existing customers than it is to bring in Nuance wait wait wait yeah um so yeah and maybe even you know AI is an opportunity for you to to drive those qbrs as well right streamline that internally that operational summarize you know tickets and do that analysis as well got me thinking now well uh my um Paul Luke our CEO is on the call please he's gonna have a long list after this tell you what um guys um I guess you know from an opportunity perspective uh we've talked a little bit about msps but focusing on the customers uh so the guys you know on the call can understand which clients might be focusing on this more are there particular industries that you've seen that are jumping on the AI Automation bandwag in more than others that if they've got them on their books that they should be jumping in having those conversations with yeah definitely interesting the the risk-averse um so I guess in the legal sector that sort of it's a hesitation between ban at all uh versus wait wait for something that's more governed like a chat GPD for Enterprise or Microsoft co-pilot or it will more easily be easy to govern rather but I think what we're seeing in terms of Industries is is those that have the combination of creativity analysis and problem solving so it'll be ones like um so media and entertainment education for sure like that definitely an opportunity and a threat like if you think about education the way that it is traditionally served up it's here is the curriculum delivered at Apex and if you're super smart you might get bored and if you're not so smart you might get left behind whereas if you've got AI That's incorporated into it has the whole curriculum in it it means that they can learn at the pace that suits them which means that everyone is going to learn better there's a few teething issues that they're going to have to work out like how do you test the students or because the traditional write an essay it doesn't really work anymore so they're talking already about more oral examinations and actually writing getting the old pen art and having written supervised examinations so they've got a lot to work through but um education and then management consulting and then um not necessarily of the the creativity but more on the analysis side financial services are taking this up in a big way so they're very interested in the likes of say the Azure open AI connector where you're actually connecting it to their data sources and um less than the um the interface of the chat gbt type interfaces in education um I'll just plug my other book chat GPT for Education um which you can find out all about that in if you'd like to it's available on Amazon it's cheap too it's only nine bucks a copy um education uh reinforcing that as well I am my wife works in the education sector and I know that there's a lot of um a lot of teachers that are developing uh you know lesson plans hey I need a lesson plan on this because I'm you know filling in um and they need to tie into these curriculum links um so it's speeding up that side of it um but I mean there's that risk of that bias like you sort of said before as well right so interesting I said I don't know if that's well it's probably not being addressed in that sense um Nikki I also mentioned that well both of you have mentioned well Matt you just mentioned Financial but um Nick you mentioned that earlier as well that your the your background is in sort of financial would you mind to touching on a little bit of where you're seeing that as an opportunity it's it's it's amazing I think I think yeah I was seeing a lot I do a lot with accountants so I'm I'm off working with the audit Association of Australia the insolvency restructuring and turn around Association I like them because their annual conferences in Queenstown um but also doing a bunch of presentations for childhood cancer signs and I think the professional bodies are really struggling with the ethics around it and they do they struggle with ethics around everything so that's fascinating but um demonstrating to accountants you know things that they that I think the issue Professional Services have is they're very used to billing by time and these are things which save time and there's this big conflict in their mind about how do I ethically Bill a client when I've now got a tool saving me time and I think that's a that's a struggle across the board and secondly if I get chat GPT to do something how do I know the results are accurate or or it's not leading them somewhere and the understanding that as a um as an organization you need to actually go and deliver um advice to your customer and it's the individual who gives the advice not the technology so they can't rely on that there's a whole bunch of ethical minefields yet there's a million tons of savings and I've got a story and I'll try and be brief because I know we're limited for time but I was at a accounting conference and somebody got on stage and said the big and this is like three years ago the biggest technological advance for accountants in the history of accounting has been the invention of laptops that run two screens so you can see one thing and copy it to the other app right that was that he was being serious and and yeah when you when you look at that and you see what you can now do with AI you understand why it could have just such a big effect in that industry so anything Financial Services is is as as Matt said taking this up in a huge way but I say anything Professional Services it applies for but we're now seeing it in manufacturing and across the board it's a horizontal solution that can apply almost anywhere hey your example of Amazon earlier is uh you know that's very large sort of manufacturer kind of thing you know um yeah interesting um Andrew how about you yeah look up I mean I I think um the most interesting thing about particularly chat gbt and more broadly AI is that they are quite a democratizing technology I think we're seeing it across the board everywhere um we're sort of seeing that traditionally more risk-averse customers are still risk-averse but uh they're being brought in it's almost a little bit like the tail wagging the dog in some cases where they're finding out the staff are actually using this and they didn't realize it so the risk-averse managers are going oh well hang on if we are already getting benefit out of it we may as well start to adopt it on a larger scale but typically our customers that are already um you know like to stay on the on the Cutting Edge attack are right into into it so I it's sort of interesting to hear the perspective from from Matt and Nick because we've sort of seen a pretty broad across-the-board Approach at the moment we don't seem to have that many Industries specifically it's more about the individual customer and whether or not they they're you know let me let me link questions do you have any industries that are very adverse to it um we don't have any industries that are very adverse we have customers that are very adverse to it and I think that's the thing everybody you know reads a lot of emails writes a lot of emails a lot of Industries write a lot of reports and do a lot of that sort of Big Data uh you know processing of lots of text I suppose and uh from a human perspective I think all of those things can be done or assisted um by gbt it just depends whether or not that's endorsed by the business or not okay so for the businesses that are the you know if you've got MSP clients that aren't and I would touch on the time thing I think Nick said your financial services that they're sort of building their time thank goodness msps have moved to an MSP model where they're not billing time for those of you that aren't breakflix because that's um straight to your bottom line right uh or you don't have that worry at all but um uh focusing on those customers that that might be more adverse to it how do you address that he's there like is there an approach that you guys have got to sort of explain the benefits we're we're actually just starting a program in conjunction with next DC because one of the big risk-averse things is there aren't any real AI posted AI Platforms in Australia so even if you use Azure you're using Azure in the US to do large language model stuff um so we're doing some work with next CC on hosting some um expensive gpus and an hpe kit to do large language model processing using the Llama models which come from meta those models in certain areas perform just as well as some of the GPT models but stay within your own environment so work on your own data in your own environment nothing goes out to the cloud that type of thing when you make that work and yeah wait to 18 months of order a GPU now wait 18 months you'll have one but but when you do that type of stuff that for a lot of organizations who have Regulatory Compliance rules so anybody in law enforcement or in investigations or in significant compliance lots of government um don't want data going outside the territory um and and you have to if you're using those models at the moment there are options and we're just starting those where you know we're running those up in local data centers um as as you know hosted environments where they can actually run it effectively on premises that's I think as as with everything Cloud's now hybrid AI is going to be hybrid as well and those models you can download and use are pretty good now especially if you fine-tune them around your use case sort of expand on that as well you know I think trust is a really important sort of word in this AI discussion I I think um AI is is such a complicated Beast of a technology and it's really amazing that the things that it can do but sorry it is really amazing things that it can achieve but um you know we really have to trust in the companies that they're doing the right thing and so um you know I think the customers that are not wanting to engage with AI on an organizational level or have you know certain reservations or legislative requirements as you sort of touched on their neck um I think we're going to have to build that trust in these AI systems because they are different they are new and they are a whole different operating model to a traditional sort of RT system so I think that the best approach there is really um you know as you said earlier Ben to get in there deep with your customers and really understand what they are trying to achieve and then whilst you're um working through these processes you can identify the processes that are most likely to be impacted by AI in a positive way and a lot of the time you don't need confidential data to help um you know with these processes it can be something small to start with and really start to build that trust up and go from there but yeah I think trust is something that takes a little bit of time um to establish for those type of customers I was going to add to that that um you know Nick you've mentioned the the opportunity going deep um if but we've also talked a little bit around the the resourcing that that you know we I think we all know we're having a challenging time here in Australia getting stuff um if you identify opportunities I'd also encourage those on the on the line to um you know partner with Partners um you know going back to what you said Andrew trust elements I think there's an opportunity there for everyone um but if you're not specializing in it to identify those as well just a little bit of tidbit we don't do AI so I'm not I'm not spooking anyone but I just think it's it is an opportunity that um I do see that you people may lose out they may lose clients as other other people come along that can do these things better and you know it's one of those it's a great industry that you have to keep learning right it's constantly challenging it's constantly evolving I do think that that AI is one of those one of those areas whether you're going as detailed as what Nick is I think you're at the very least um you know you need to have a good understanding of so you can have those business conversations um so with that in mind uh because I am being told that you know that we've only got nine minutes left um like to sort of get your magical you know idea of where this is all going right maybe maybe someone can ask the AI and tell us what that Ed thing statistically speaking but um where do you think we're headed um this has evolved very quickly I remember I remember when the first time I um I punched I was writing books for my my daughter with you know all family members and things like that and it was blowing my mind absolutely blowing my mind a bit like BR did um but um where do you guys think that this is heading in a year maybe two years I think it'll just be ubiquitous it'll be part of every everything that we do and it won't it the the novelty will wear off in a little bit it'll be like the internet or like mobile I think it has like a similar level of impact in that it's just going to be something that everyone uses and um yeah so yeah Nick you mentioned a law enforcement before if you say Minority Report I'm leaving well actually if you read the terms and conditions for chat GPT you're actually not allowed to predict crime with it just just so you know really yeah that's one of the terms um but and there's a couple of things and one of the big things which is going to shift this is going to be regulation so if you're in New South Wales today the New South Wales government has opened for submissions on their AI platform and that's open till the 20th of October so if you want to say at least on how government sees it that's something we should do so there's going to be a whole heap of Regulation up here some of that will be good some of that will be bad that's going to change where we are I think Matt's right it's going to be ubiquitous it's going to be everywhere but I think we're also going to see what we saw when we shifted to cloud and what we saw when we shifted to distributed computing from Mainframe and mid-range is that a whole bunch of household names in the tech industry are just going to disappear and there's going to be a whole bunch of new ones and we see that every time that and I'm living through my third big change like this every time there's a big change like this you see quite a quite a significant shift who'd have thought 20 years ago that AWS would be a thing right so so there's or even Google to be honest with you but there's there's those shifts happen so I think we'll see that happening again which makes it a fascinating time at the moment to go make lots of money because it's new and not a lot of people know how to do it again in a few years time it will be ubiquitous it will be regulated it'll be very different so the opportunity is now um Andrew anything absolutely yeah so so I I think um you know I think Matt and Nick have some summarized it really nicely I think the only thing I would add is that I think it's going to raise the Baseline a little bit so the the sort of example that um I'm totally stealing this from a podcast and I can't remember who it was talking about it but somebody was talking about the use of AI in education and they were sort of saying that like uh it just raises the Baseline it means that you have to be at least as good as the AI because everyone has access to this tool and and everyone can and can make it right an essay and so it just raises that bass line and says right here's the tool if you can't do something as good as this then you know that you know that is that is that Baseline and I think we're going to see that across the board in a lot a lot of things I also sort of Wonder um you know maybe it's a little bit more philosophical than technical but I do wonder whether it's going to change the way that we communicate I think that emails are out of control I think teams messages are out of control I think that uh you know we already have lots of strategies in place to try and uh sort of triage the way that we interact with each other and I I wonder whether um AI is fundamentally going to change that um because it will really be my AI talking to your AI rather than you and I having a conversation unless we pick up the phone and and in some aspects that wouldn't necessarily be such a bad thing in other ways it it may not be so good I don't know so I I really wonder about that but I don't think there's any question that this is going to be a ubiquitous piece of technology G I think it's really going to raise that Baseline and it's going to give people a lot of opportunity um which is really really cool cool guys we're at five minutes time so final question um what are two things from each of you that the audience should think about and go and do as soon as they leave this webinar if they're hearing this as a podcast I'll let you guys think about that in a second but if you aren't or uh if you aren't already using it extensively a really simple thing to do is set chat gbt as your home page and what that does is it it means that you don't fall into the Trap that many do which is uh that they use it for about five days and then kind of forget that it's there and remember sporadically but if you've got it as your home page then as opposed to Googling things trying it out for different use cases and the other is to for for a team I'd suggest that managers give their team half a day and say automate the crap out of your job just think about all of the things that you do and dedicate the time to test where what is it good for and what is it not good for um and uh also switch off that button that makes your your data go into the model Nick Andrew Phillips oh I I I'd say first of all um what chat GPT is lovely and it solves a whole bunch of problems there's another tool out there which I really like called perplexity so perplexity.ai um that integrates with a whole bunch of different things including chat GPT but the thing it's got is it's got access to the internet so when you ask it a question it browses the net it gives you citations the way it found it it gets over the fact that those models have don't have modern up-to-date data I use it instead of Google Now um it's become my go-to tool it sits on my phone it sits on my PC it's really really cool so go and play with perplexity.ai as well as chat GPT um just put that in there yep I will do my to-do list I I shall put that in the chat shortly talking first then I'll put it in the chat I can't multitask I'm a guy go for a walk while you're talking Nick we'll just walk over here and do this yeah but anyway so perplexity is really cool the other thing is get involved in automation now whether you use power automate or whether you download my free tool workflows um there's a bunch of tools out there that allow you to integrate AI into business process Automation and start experimenting with things in your business go automate a tiny process today do those two things and I think you're setting yourself up for success in this space you can then lead your team to do cool stuff love it Andrew awesome thank you I feel like somebody needs to say that that you know one of the things I'd like to do after this meeting is go read Nick's book because that sounds fantastic um I think uh yeah that's right that's why I go around my own um I think uh as a Microsoft uh you know team we sort of have to um plug cover pilot I think for all the msps out there that interact with Microsoft 365 and are looking to expand that or continue that practice definitely do some reading about co-pilot I mean there's more information being provided every day I think this this is going to be something that as msps we really need to be on top of as this starts to come out because all of a sudden customers are going to have uh more tools available to them than than they have in the past to do some really amazing stuff with and I think we're going to get an awful lot of questions as an industry about how do we best interact with that and what are the security things um yeah I think other than that just um get in and play with chatbt really I think you guys have covered everything Matt talked about the security aspect so um yeah I think just play with it and see what it can do for you great guys that was um that was a great conversation I really enjoyed that I've actually taken a whole page of notes um of bits and pieces that I'm gonna you know okay certainly take away uh and I know that a few people will as well uh so a huge thank you both from me and everyone on the call um for those of you on the call uh this will be recorded uh so we will be sending it out so if you do want to share it around or if you want to refer back to it um in a couple of days we'll um we'll have that out to you it'll like I said earlier it will be available as a podcast uh and finally uh if uh you do want to see any of our previous msps in conversation or our up and coming ones check out uh hostednetwork.com.u forward slash event because that's where the recording will also be um guys again finally again thank you so much I really enjoyed the conversation so much thank you so much for all of your time today and everyone for putting aside an hour to listen to us ramble on about AI for an hour uh thanks guys we'll leave it there