How AI Will Affect Your Business in 2024: Insights from CROFTI Leaders

We’re thrilled to welcome new faces to the REDD studio for our final episode of 2023! In episode 45 we are joined by Lucas Meadowcroft and Dean Cavanagh, tech entrepreneurs behind CROFTI and Tribu. We delve into how AI will affect your business in 2024 and the complexities of AI integration, particularly in the MSP realm. They candidly share their experiences, highlighting the challenges faced when introducing AI automation to MSPs, with a primary hurdle being the hesitance rooted in accuracy concerns and resistance to change. 

Their fundraising journey sheds light on the evolving investment landscape, where VCs now demand a robust product-market fit, emphasising a substantial monthly recurring revenue of $100,000 before considering funding—a stark shift from previous norms. Looking ahead, the duo discusses Tribu’s strategic pivot towards vendor integration, aiming to streamline AI adoption within the MSP sector. 

This dynamic discussion showcases the nuanced hurdles encountered by startups in embracing AI. It raises critical points about accuracy concerns and the evolving expectations of VCs. As AI continues to transform businesses, Tribu’s focus on vendor integration signals a proactive stance, recognising the pivotal role AI will play in reshaping the future of MSP operations. Join the conversation for insights into AI’s impact on business and the strategic approaches needed for successful integration. 

#TechEntrepreneurs #AIIntegration #AIAutomationChallenges #VendorIntegration  #BusinessTransformation

00:00 – Opener
00:25 – Guest Introduction
01:23 – Lucas’ Career background
06:05 – Dean’s Career Background
08:20 – AI Integration in Business Processes
13:14 – Multimodal AI and Future Expansions
16:23 – Business Roadblocks for Leveraging Automation and AI
21:53 – Exploration of AI Maturity and Tools like CoPilot
32:05 – Obstacles and Learnings from AI Sales Journey
38:12 – The Shift in Angel Investing Landscape
41:21 – Focus on Vendor Integration
45:01 – Closing Thoughts and Remarks
45:19 – Outro

If you would like to discuss any of the topics discussed in this episode further with a REDD expert or if you would like to be a guest on the show, please get in touch either via our website, [email protected], or through any of the links below. https://redd.com.au

https://www.linkedin.com/company/redd-digital/
https://www.linkedin.com/in/jacksonpbarnes/
https://www.linkedin.com/in/bradley-ferris
https://www.linkedin.com/in/lucasmeadowcroft/
https://www.linkedin.com/in/dean-cavanagh/

(00:20):

Hello and welcome to Redd’s Business and Technology Podcast. I’m your host, Jackson Barnes. I’m your

(00:24):

Co-host, Brad

(00:25):

Ferris. Today we’re sitting down with two guests, Lucas Meadowcroft and Dean Cavanagh. Thanks for coming in. Really appreciate your time. We’re speaking about the journey through the business and technology world that you’ve gone through, AI and everything happening in the last 12 months and the future 12 months. So looking forward to getting some your insights. Maybe if you want to start Lucas with introduction to who you are, give me the quick version, maybe the one minute version. Then we’ll jump over to Dean.

(00:48):

Yeah, thanks for having me. Next guys, Lucas Meadowcroft, CEO co-founder of a few tech businesses. In this case, our Crofti and Tribu, which is what we’re talking about today, which is all things AI.

(00:57):

Hey guys, Dean Cavanagh, I’m CTO for Crofti and for Tribu, one of the co-founders there and what else are we talking about? And that’s it. That’s

(01:07):

Good. We’ll just be throwing questions. Jason mate, do you want to get some insights around ai? We’ve been

(01:11):

Just crushing the intros today. Just for everyone listening, this is a third attempt at it.

(01:15):

Nailed it.

(01:16):

Alright, let’s kick off. Maybe Lucas, we want to go through your business journey. Obviously we know a little bit about the background, but maybe if you introduce your journey to get to where you are now.

(01:24):

Yeah, absolutely. So MSP is my complete world for the last 20 years. MSP being the mini service provider, essentially what Red provides on a daily basis, that journey really quickly evolved from being an IT engineer on the road going and servicing clients to doing more account management sales through to then running an MSP and focusing on the business world. So I was running into the business world IE running the business 15 years ago and then had to learn all things business and not MSP and not technical. What I found throughout that journey was, I guess what I loved is listening to customers about actually what they wanted to solve in their business. And what I realised really quickly was IT support was essentially just something that they already expected. It’s like what the next step was. So looking at innovation, looking at back then it was like very Excel spreadsheet process improvement. How do we get the information out of our spreadsheets and what can we do with that data? And obviously over the last 15 years it’s evolved from implementing all the systems you can think of on the planet. I think we’ve implemented over a thousand different systems now, and then the last four years has been very much AI business process automation driven.

(02:41):

So you went from essentially tech to more account management and then what you were just passionate about was the innovation that technology had into businesses and that’s why you essentially because you exited the MSP and then do this full-time now, right?

(02:54):

Yeah, absolutely. Yeah. So we exited the Crofti IT support MSP two years ago. We went all in on AI from a Tribu standpoint, our Tribu organisation four years ago, and we’re providing AI consulting to organisations around the country and even overseas.

(03:10):

Cool. Let’s go maybe the Tribu story first, then we’ll jump over to Dean.

(03:15):

Yeah, absolutely. Well, it’s a good segue anyway. This is essentially his baby. So the biggest pain points we had in our MSP was essentially why we started Tribu. So what we were already trying to solve, manually solve with the Excel spreadsheets and exporting data around the PSAs and all that. Dean dive into that bit further. That’s how the Tribu journey started. But what we realised quickly over that, say first six to 12 months, it wasn’t just us that had the problem, it was every MSP we spoke to. So anytime you go to a MSP conference and you have the conversation about what you’re doing in your MSP, and then we started having a conversation around, well this is what we’re trying to do, this is what we’re trying to solve. They said, oh, could you solve it for me too? So it just evolved from solving it for us to solving it for others. So they did dive into the technical side. What was

(04:00):

That problem? We were targeting the triage. So as tickets come in, getting ’em categorised correctly, getting ’em to the right technicians, the right board or queue, depending on what system you’re in, that sort of initial, let’s deal with the ticket upfront and get it to the right people, right place, the right category. So scheduling, that’s a big part of getting it to the right schedule.

(04:23):

You have to tell the Excel spreadsheet story working with technicians and engineers.

(04:28):

So I guess that comes down to prioritising tickets and helping engineers. I guess when the technicians, if they’re starting to get overwhelmed, say they’ve got 20 odd 25 tickets in their ticket queue to help out various customers across a lot of different clients, a lot of ’em just get to a point where they bury the head and the sand, I dunno where to start. They’re either start cherry picking clients that they like the most maybe or the problem that’s easiest to solve, whatever, instead of doing what they should be doing now. So we used to export it out to a spreadsheet. Each technician had their own spreadsheet and they’d have to go through start of the day and go, right, this is what you’re doing in the order that you’re doing it in and deal with it that way. So that’s part of the prioritisation that came out of it as well. Yeah.

(05:11):

Wow. So literally sitting there so the technician rocks up in the morning, they can say, I have to do this job and then I have to do this one and then I have to do this one. And they couldn’t move on to the next one until they did that job. So very much handholding, but we found

(05:22):

When super micromanagement, super

(05:24):

Micromanagement do not recommend, but I guess because dean’s very technical, what I noticed throughout that journey is the output we got from technicians and also the happiness we got from the techs as well or engineers because they love structure, they want to be told what to do and so they were happy and we got better outputs because I know Dean wasn’t happy doing the manual export job.

(05:48):

That’s three hours of my life that

(05:50):

Every single day I can get back. But it was a great outcome for the company and so we went, okay, well this can’t be a manual job for Dean for the next 10 years. Then how do we turn that into an actual product to

(05:59):

Solve? No problem. Makes sense. Let’s go your background. What did you do as you were a technician and an MSP before you doing this or what were you

(06:07):

Doing? Yeah, it was a fairly, I guess common sort of path starting out as just a technician and in the early days, so 23 years I’ve been in this industry and across technology in the early days. I was actually Lucas’s manager back when he started, so I’ve got three years on him on this. But just go through just working with small business largely and a few home consumers as well in the very early days, which I won’t go back to

(06:33):

Teaching people how to use Facebook for the first

(06:35):

Time, all that sort of fun stuff. And then moving more into business projects and I guess that sort of project focus became a big part of it. Understanding scope, understanding outcomes from a business perspective and a total strategy for a company. And quite often realising that most customers dunno what they want. They know they want an outcome, they know they want to change something, but they dunno where to start. And that’s really where a lot of this fostered.

(07:01):

Typically what’s better if some potential customer comes to you with the problem they’re facing more so than what they want Otherwise that’s towards a bad outcome. But let’s keen on pinging your brains on AI and what’s happening in that industry. I feel like there’s a lot of hype around it now and that kind of thing, and it’s almost a bit ridiculous in terms of what AI is because all these marketing companies just reba any level of technology automation as AI these days. Maybe Dean, if you want to go, what is AI in your words?

(07:27):

So I guess I won’t go too technical into it, but essentially it’s mimicking our decision-making as humans, our decision-making skills, taking from what we’ve learned, our experiences and then being able to go, well, this is most likely this. And AI is largely around probability to do things or whatever the answer should be, and that’s essentially what we’re doing as humans. We might think we know the answer a hundred percent, but there’s always just this sliver of doubt, I guess behind things often and that’s essentially what AI is as well. That’s important for everyone to consider. It’s never going to be a hundred percent right. Humans are never a hundred percent, so can’t expect it to be the same in the machine.

(08:11):

And maybe anyone want to answer this one, how has AI in the last 12 months changed the business world? Maybe an example of something AI that you’ve put into a business, for example.

(08:22):

I think the biggest thing that’s changed that we can have a conversation with someone about AI now. So for three years previously we were hitting our heads against the brick wall. We could not talk to anyone about it, no one would listen to us. That was our entire world for a long time and it’s like I say, a long time, long time in the tech space and we just couldn’t talk to anyone. It was really difficult to have that conversation. So is

(08:43):

That basic

(08:43):

Of

(08:43):

The back of chat GPTI

(08:44):

Guess? Yep. Chat GPT launches within a few weeks. People now know what AI is and next minute the only conversation we’re having is about ai, is about ai. So that’s I guess from a personal journey because even in the early days when we raised capital from VCs in the us, even having an AI conversation, then it was still very much an unknown conversation. So it was like, oh, I think it’s this, and actually every investor thought it was something different. So even that was a hard conversation compared to these days. From a business perspective, AI can be used utilising across the entire board when it comes to your organisation, whether it’s marketing, sales, operations, service delivery or product development, and even HR people and culture finance, it can be utilised in every aspect of the business. The biggest disconnect I find in the conversation we have these days when it comes to AI is, well, we just need ai. What does that mean? And so we always talk about this, it’s like 95% of what actually what we do is automation. 5%, which is what Dean was talking about, the decision making 5% is ai, 95% is just pure automation, which we’ve been doing for 20 years. So that’s the interesting conversations that we’re educating people on what AI is these days. We’ve

(10:01):

Definitely seen that. And like I said, everything’s as AI these days. Have you got an example maybe the scene field seeing on what you’ve done, maybe they call AI project for a business locally?

(10:13):

Yeah, this is Dean’s entire world. I would probably pick the business insights, just be able to extract data and forecast data that the customers didn’t know where their business were previously.

(10:24):

There was a, okay, so for a couple of examples I guess one was for an energy company and we uncovered something that they didn’t necessarily understand from their data. Now this isn’t a huge amount of AI in it necessarily, but it lets them forecast into it. But essentially we pick out their contracts essentially from the time that they get paid versus when the job is finished and it was like 40 days as a gap and it varied a lot. The value of the contract didn’t matter. It could be for a thousand dollars contract or a $4,000 contract. It didn’t really matter What it turned out to be is the reason that it could vary from two days from payment or 40 days to get payment was purely because there was a report that needed to be written in between before payment could happen. So the sooner the report got written, the sooner that sort of cashflow would tighten right up and they weren’t able to visualise that or see that and then we could start to measure that and forecast, okay, well this is what your cashflow is going to start to look like as you bring this closer.

(11:25):

So that’s the insights side of things. In terms of some of the actual, I guess AI predictions, it’s largely been around, say email for example, being able to classify what has actually been coming into a group email inbox. This is partly where trivia sits as well that’s very focused on IT support, but this was just a general, all emails received at this company all went through one mailbox, pretty interesting sort of strategy, but it is what it is. They thought everything was related to getting quotes and they were getting, let’s say 200 emails a day. They thought it was almost all quotes. So we were working around some automation on that. What we found out was when we start to classify those emails for them, it turns out that it was only about 30% as quotes, and the rest of it is just general inquiries, invoice follow-ups, things like that. And then we realised, okay, once you can see that and it’s automatically categorising it for you, you can start to target, well actually if we can take this other quarter to a third of your emails and automate some of the responses back with the general information that they want, that’s a third of the workload. And this is 12 staff manning an inbox that’s a third of that workload taken off their shoulders straight off the bat. And

(12:44):

You built the solution to

(12:46):

Yeah, we built, yeah, so we built out the categorization to be able to build that out. We hadn’t got to how we automate some of the q and a, the q and a side of it, and I dunno how much you want to go down on this path, but a lot of that, being able to answer things properly and the way that the company wants to, requires the company to have good data themselves and that’s where things can get a bit

(13:06):

Tricky. That’s probably a good segue. So I was going to ask, one of the questions is how do businesses now get their data ready for AI in the future? Even now you’ve got Adobe and all these other Canva and all has the AI add-ons, right? But if you don’t have the data in the right section and it’s already in the right kind of format, you can’t really leverage it. What advice would you have for businesses to get ready to leverage AI in the future? I

(13:31):

Think a good workplace aside is actually structured versus unstructured data.

(13:35):

Okay, so I guess around structured data, which is it has labels, it’s already got labels in it. If we take for example, well, I’ll pick on you guys with your tickets, you’ve had a bunch of support requests come in over time, you guys have labels for them. This is a support query about email, whatever it is, things like that. That’s essentially you’ve structured that data, we know what this looks like and it equals that unstructured is where, hey, we’ve just got this mess of documents could contain anything. I dunno what it is that’s unstructured. So when we build models around that, we dunno what the result is. All we can say is, Hey, we’ve created this model around a bunch of documents, this new one that comes in, it looks like this group of documents. Go and look at that group and you’ve got a match.

(14:24):

If you think about it, it’s like a persona for a document or whatever the case may be. So it helps in that sort of space. You still need to categorise it at some point. So for when it comes time for how companies can implement this, it is largely around creating good, typically good processes upfront to understand what the process is for, whether it’s whatever the workload is, whether it’s production, office, administration, that sort of stuff. To have a good process in place when this equals this, do this thing when I see this, but it also has this other thing in it, it’s going to go a slightly different way, whatever the case may be. It’s a lot of business rules, but typically even if it’s just a human learned behaviour and it’s gone a particular path, that’s where I can pick up and say, okay, well this content moved along this path. I’m going to move it along for

(15:13):

It. So your advice there is to essentially evaluate where information’s coming into the business and make sure it’s categorised and set up in a way where in the future you can potentially automate

(15:23):

That, that will get you the cleanest results. Typically, if you don’t have that, and I know that can be hard for a lot of businesses to set that up, the next thing I guess I would focus on is at least some level of consistency. Make sure that the way things happen in the business remain consistent for a period of time. It might be you might have to micromanage a little bit to make that happen, but then you can at least pick that up even if it’s for a few months, whatever the case may be. You’ve got a few months of good behaviour in a process that the AI can then learn from and start to replicate for you. I assume

(15:57):

Also if you’ve got your data set up like that, when you do come down the road to do some sort of analytics or inside or automation projects can be a lot cheaper.

(16:04):

Yeah, absolutely. It makes it so much easier. So there’s

(16:07):

Probably some incentive

(16:08):

Right

(16:08):

There. Absolutely.

(16:10):

We’re seeing that as a blocker sometimes when you put automation in place and the business is like, we want to automate this. Then you have a look at it and it’s just not on the right format. Therefore someone internally has got to spend weeks going and getting it done properly and then it just doesn’t happen. That’s right. It’s a of a problem. What other roadblocks have you seen from businesses not being able to leverage automation or AI because of the other issue?

(16:32):

Typically it’s been around, we’ll go controversial and say people that don’t want to adopt it or people that will throw a spanner in the works, they think they’re under threat, I guess their role, whatever the case may be. And also around accuracy. A lot of people get worried about accuracy. They’re probably the two big roadblocks when you’ve got the data thing sorted. In a good process it’s going to be people and accuracy and understanding really where accuracy sits. So I dunno if you want to dive in on either of those two points.

(17:02):

Well, how do you get past those blockers around the people piece?

(17:05):

It’s got to come from leadership. The companies who we’ve been working with that have adopted not just AI but just adopted the future of working, it’s their mindset. Their mindset is they want to move forward, they want to be better, they want to be better than their competitor. They want to be the leader in their industry. That mindset and drive then from a leader point of view then fills down to the rest of the team members and whether the team members that don’t embrace it are there anymore or not, that’s a different conversation, but it’s coming from the leader and it’s coming from the top.

(17:35):

I will say on that side of it, I guess if you can show that one leadership is using it as well, but also to say, Hey, this isn’t a tool to replace you necessarily, and quite often AI is not going to replace the entire workforce. That seems be a nice big hype train that’s out there at the moment. It’s going to take a long time before in that sort of space that’s not like the next 12 months and I see a few speakers about AI get up on stages and say, Hey, everybody’s under threat. It’s a big danger. It’s not around the corner. It’s coming eventually and I don’t think it’s going to be as big as what people are saying. That’s just my personal opinion out there. But essentially if you let your staff embrace it and show them how actually this elevates your career, you can get rid of the stuff you don’t like doing and focus on the thing that you actually like doing because the AI is going to do this boring bit over here. If you happen to the boring bit, that might be a bit hard, but this normally is a career elevation for a lot of people. If they embrace it,

(18:34):

It’s a good way of looking at it. It’s probably something you do pretty well, Brad. He’s always saying ttts thing.

(18:40):

Yeah, that sounds good.

(18:41):

It’s going to give you a good base to start. You can save however much time depending on what you’re doing. I find I always have to go back in and edit, but it just takes away a lot of the, you don’t have to deal with a blank page anymore. Get something to manipulate and work with straight away.

(18:57):

Things that would take you hours to do is now literally five minutes. What would you not want to make your life better? A real quick example off the back of what Dean said is quick wins, what we call quick wins. How do you get leadership? How do you get staff members? How do you get anyone in the company to embrace change essentially AI automation, quick wins. So quick wins could be as simple as example of the other day was like staff members looking at a PDF document taking information from that PDF document and putting it into a system. That’s what they do on a daily basis. Is it a full-time job? No, but it’s a part of the job. So rather than doing that manually, you extract the data using AI and you input the data into the system so it’s automatic. You don’t have to manually input that data anymore. So just if you can show someone a quick win of how AI can be used in a business sense, then they go, oh, what else can I do? Then it opens their mind. They don’t think simple things like that. Don’t realise that that is even possible.

(19:52):

Good advice and it’s often just then redeploying that person to do something else that’s more meaningful.

(19:58):

What would you say to people who, I think you probably get this a lot of the time, so I’m going to ask it and get your response out of it. Say it’s accounts payable automation and know us, it’s big numbers. If the AI changes numbers here, it could be backwards heaps or payable wrong amounts for example. Do you still get pushback in that area or people okay with

(20:18):

That? I think that’s the accuracy conversation. It

(20:20):

Is a bit. Look when it comes to hard numbers like that, and I will say a lot of the LLMs that are out there, chat GPTs probably the most common one hasn’t got that right a hundred percent of the time and people need that part to be right 100% of the time. So what do you do for those sort of things? You don’t have that decision making done by the ai. That’s the short answer.

(20:41):

So it was more sensitive information, whether it’s dollars or whatever else. If you’re leading AI to deal with that and if there’s a mistake, it can have a massive business impact if it’s wrong, which basically is accuracy. So your advice is to not automate those things.

(20:58):

You can automate the creation, but I’ll put it to you guys. If you guys happen to have a large tender that’s about to go out, for example, I’m sure more than one person looks at the final result that just remains in place. It’s just the initial build and all the time that goes into building that up can be done by AI and then you’re going to have some fact-checkers, let’s call it on the tail end of it. That’s already probably part of your process for critical information to just do a check that still happens and that would remain in place.

(21:27):

Someone told me yesterday that it’s actually an AI tool, I can’t remember the name of it, that can actually help you with tender responses and just do 80% of the work. I was interesting

(21:37):

Out of the box today, you can do ones that we’re literally quoting on right now for clients is to automate their entire service reporting. So something that would normally take one or two full days work to write a full service report we can do in minutes.

(21:51):

Yeah. So what’s the majority of the work that you are doing for businesses, processes or functions you are putting AI into?

(21:58):

Yeah, I guess our background is all about business process internally. So we’re not going out there and creating new AI apps and helping people solve world’s problems. It’s really focused on the internal automation aspect. So how do we remove all the mundane tasks from staff members and give them time back like Brad said before, so they can do more meaningful work. That’s our core aspect.

(22:23):

So that’s service report example, so you’d be pulling from multiple

(22:26):

Systems

(22:27):

And data sources?

(22:28):

Yep, multiple data sources, multiple systems. So essentially a person who needs to put example the other day was in the NDS space community worker. They would pull in multiple different data information from a multiple different systems they use in their business or sorry, they wouldn’t pull it. They would look at those systems, they might export ’em to spreadsheets or they might just look at the actual application, but when they’re writing their report, which might be a 10 or 15 page report, they are just pulling that information from their systems and writing the report. We can automatically pull the data from the systems. We can use something like touch t, it’s different technology, different backend, but would use a similar technology, look at the information, automatically create the report, get it to the final point that then the person is checking it to make sure they’re a hundred and happy with it and then they can send it out. So

(23:20):

You’d pull all the data, you’re effectively creating a new data set from a bunch of disparate data sets. You create a new data set, then you have an LLM or something, look at that new data set and come with some sort of insight and then you’ll have the third person, the reviewer, then check if that insight makes sense

(23:40):

Essentially. Yeah, that’s good. That’s a big part of it and I guess talking because spoken a bit about chat, GPT here and something like that. Like Lucas mentioned, NIS, there’s obviously sensitive information about people in that sense. It’s not something that we would recommend goes to chat GPT, that’s an open sort of, it’s offshore, all that sort of stuff. People will ask questions about security and that you can take these LLMs and make them your own in your own hosting your own sort of data centre, whatever the case may be so that it remains yours and it’s not third parties necessarily weighing in on it or maybe leveraging whatever’s happening through it and having access to that data. So that is still there. I guess I bring that up because a lot of people go, oh wait a minute, does it mean I need to give everything to open ai? No, you don’t. Open AI is just a great tool to go, is this in the right space to be able to do this thing? Yes, yes it is. Okay, let’s look at bringing this in-house

(24:34):

And out of the box. If most, I say 80% of organisations have Microsoft 365 now on the SSM E space, you can do it within 3 6 5 ecosystem. As long as you’ve ticked the box to say you want your data to be stored, say in Australia, then you can go ahead and this can happen immediately.

(24:49):

What’s your thoughts on copilot so far?

(24:51):

It’ll be great once it’s publicly released to the SME world. We haven’t doubled it too much. We’ve already been doing what copilot can do behind the scenes using Power Automate. So from a user interface and a user in a business using copilot on a daily basis. We haven’t even dove into that yet until it’s more publicly about the enterprise side.

(25:14):

I have a lot of technical questions. That’s it behind it. I’m like, okay, so they’re saying it’s doing all this. I wonder how well is it versus Hey, is it just an open AI plugin of sorts that has a nice little Microsoft badge to it or is this very personalised to the business? I know from others that I’ve spoken to that are using it, I haven’t had a lot of hands on time with it. They’re saying, no, it’s personalised, but it’s still not quite right. I guess I do want to dive into it with my technical hat

(25:42):

Because it can’t be rolled out to SMEs per user, per month scenario unless you go and buy the enterprise licences right now, it’s like, well, you can do everything that it says it can do no documentation on copilot. You can do everything out it as it would do behind the scenes using Power Automate anyway.

(26:01):

Yeah, it looks cool though, but it does.

(26:02):

It does. Yeah. Yeah, it does. Early next year, hopefully Q1. Hopefully it’ll be, I don’t know, unless you guys have some insights roll out and we can publicly start using it every day, that’d be amazing.

(26:13):

Yeah, we’ll see what happens. Definitely exciting space and I think it’ll have a really positive impact on small businesses if it does get rolled out, but we’ll see what happens there. Move on. For you, Dean, what’s going to happen in the AI space next year in your thoughts?

(26:28):

We’re seeing a fair bit around this multimodal, so being able to deal with text as well as images and files and things like that. I think that’s probably going to expand out a bit more when we talk about what’s happening in, I guess the public realm. So the open AI and now with Gemini, which I think just dropped yesterday essentially, which is looking pretty impressive.

(26:49):

What’s that? So that’s chatGPT Yeah,

(26:53):

It’s like the next level of their bard, which was their original release.

(26:57):

What a

(26:57):

Name. I know Bard. Yep. That got me excited.

(27:03):

Gemini is much better.

(27:05):

My thought is that multimodal stuff will probably expand out. I know they’re trying to do stuff around sound and music and things like that, but I dunno where that will fly so to speak. Look, I think there’ll be a lot more focus on how do we make this usable to a wider range of people. I still don’t think that there’s a huge amount that will happen direct to businesses as in there’s not going to be an open AI created plugin into whatever your current ERP is, for example. It’ll be up to the ERPs to create those. So a lot of that will release I

(27:37):

Think. Well, you can save time on documents. It’s still a little bit gimmicky. It is. It’s useful. That’s why I like the prospects of copilot just because my understanding is just in 3 6 5 and that’s not all we use, but a lot of our information is in 3 6 5. But for me being able to question what’s going on in my environment and have we met with this client or show me anything around this document or what’s Jackson been up to or whatever, I like that it will be able to give those somewhat, for lack of a better term, gimmicky insights but still quite useful. It would still save me time and give me a different perspective, but yes. Is it solving all the world’s problems? Not really, no. But it’s going to be useful. Yeah,

(28:17):

I would say that we’re going to see a bit more maturity, like you mentioned before, Canva and a few of those that have AI built into them, they’ll start to probably mature a bit more. They’ll actually get a lot better at what they’re doing now. You can see where they’re headed. 2024 will be a bit of maturity around that I believe. I dunno how many massive sort of insights there are that will really big leaps forward that we’ll see. It is really hard to tell though because it has been moving at such a pace now that it’s just getting a little bit

(28:47):

Insane. I think that is part of the problem. If I just think about cha e, BT you can only give it so much to look at. Whereas like that reporting example we mentioned before, if you can connect all these different data sources, pull out that data and then give that new dataset to an AI to then, okay, look at this. So you’ve done a bit of that. Like you said, the 95%, 5%, that’s pretty powerful and that’s where you can get some real meaningful

(29:13):

Results. So business you going to fly. Yeah, I guess for the audience listening is if you have an AI first mindset, so anytime I look at a new product or using any of the tools internally, I go, well what can I do from an AI point of view before I think about what can I do out the box example? I said yesterday we met with a client, I had to create process maps and so I’ve been creating process maps for 10 years and so you normally sit down with a client document everything that they do in their processes, the accounts payable accounts are receivable, documented all their processes. Then you go away and spend an hour or two hours creating process maps on the entire process. I said no AI first mindset sat with the client. I said, do you mind if I do this with you? So I documented their process sitting with ’em and at the same time I was documenting, I put it directly into Fig Jam, which is the Figma process map version into the new AI tool and we created the process maps there and then on the spot within five minutes the whole process map start to finish completely created for me. So

(30:12):

It got in it fig jam. I know Figma

(30:15):

Fig Jam is a process map tool and

(30:17):

They’ve got some ai. So how did it save you the time?

(30:20):

Because it created the entire process map for me. Didn’t have to piece together

(30:24):

All the bits and pieces. You just voice recorded

(30:26):

And I typed, you couldn’t do voice recording. I just typed out the process. So when this task comes in, what happens after that is I then have to input that into the system. I then have to notify this person that I’ve inputted it. I then have to tell this person the next step, literally the step-by-step. You still got to get out of the person every step they go through. But once you document that inside the system and I’m just using Figma or Fig Jam as an example because it’s a good tool, normally you’d physically go and create the diagram. So create all the different shapes you’d use in a process map and then you’d spend the next couple of hours trying to figure out if that sits right, move that arrow here, push this arrow here. It takes time. Right? A hundred percent. I was not expecting it to be because a brand new release just a couple of days ago, I was not expecting it to be perfect. It was so perfect. I got the custom to review it a few times and they said it’s exactly what we go through.

(31:19):

Yeah, I think you’re right

(31:20):

On the spot with the customer in their office. Yeah,

(31:22):

It’s definitely going to mature I think next year how AI in the tools as well that are being leveraged, that’s going to mature as well. You look even prior to CHATT and stuff, there was things like, was it Google who had that phone thing where you could call the AI assistant or whatever, five, six years ago. Yeah, that’s what I was going to say. That was a long time ago. But even today, no small businesses using that. Everyone has standard IVR stuff calling through, so it’s probably going to mature and push out all those little things where you could leverage ai but it’ll get more products around that kind of space and I think that’s definitely good advice and where that’s going. Let’s pivot a little bit, maybe look us back to you raising capital in Australia as a kind of tech startup. You want to explain your journey and process you went through and maybe lessons learned?

(32:06):

Yeah, absolutely. So when we came up with the concept, well Dean was able to build the concept via spreadsheets and then we wanted to turn into a proctor. We spoke to many MSPs and they also wanted the idea. We started diving into the world of well, do we want to bootstrap this? We have the last 10 years in our businesses or do we want to get funding and go quicker? And we knew from an AI point of view we needed funds to do that. And so that’s why we made decision, collective decision as a team. Let’s go our networks, find out where we start and let’s raise capital. This is completely not our world. So we’re talking about we’re in MSP, we’re building a AI tech platform. Where the hell do we start from this point of view? So reached out to all our networks. They said we should talk to Steve Baxter, the local entrepreneur and Brissy. So we pitched to him, I probably have a post I could even share with you many how did

(33:00):

Get to in front of Steve backs

(33:02):

Accuracy. So he was running River City labs and so that was an incubator for tech startups and so we were going along to those events and we were chatting with him at one of the events. Actually, do you want to tell ’em the answer on what Steve said to us when we fixed the idea

(33:19):

About you should outsource that, you should look at Arizona for hiring staff to power that sort of behind the scenes stuff. We were looking at I guess trying to do this triage, this level one triage sort of stuff with human resources. So we were looking at Philippines initially and he said no, how about you have a look at Arizona Price is much the same and it’s going to be the accent. I guess particularly for the larger market of the us, they’ll be more receptive.

(33:53):

Arizona is the biggest per capita Philippines society in us. There you go. I didn’t know. I

(34:02):

Didn’t know that at the time.

(34:04):

And so he is like, oh no, you don’t need to build a tech platform to do this, solve this problem. Just go and hire staff. And I was like, this is from Steve Baxter. But it was our first I guess inkling to, okay, now we want to raise capital. At that point in time we had a client that we had built started building an app for before. They didn’t want to go any further with it. It’s a whole different conversation and they were angel investors. So we were chatting with Steve, Steve Best is his name, and he ended up becoming an advisor on our team and he is like, don’t raise capital here in Australia. Don’t even bother. Go straight to the us. He had not had previously been on one of Steve B’s tours to the US and met Jason Calas and a lot of the other angel investors in the us. And so Steve Best that is just pivot us directly to that approach. Introduce bit a

(34:58):

Shame he can’t in Australia, but anyway here, go

(35:00):

On. Two main reasons for this is a longer conversation and probably what we have now, but two main reasons for not raising capital in Australia. One, the requirements that Australian VCs or angel investors want you to meet is completely different to what the US ecosystem is. And so we had an idea, we didn’t really have a product. We kind of had a bit of an MVP that we put together, but it wasn’t really a fully blown out product that we had built, launching customers, paying customers, like hundreds of paying customers. So we hadn’t gotten to that point. The Aussie investor network would like to see that. They would like to see you

(35:38):

With a Polish product

(35:39):

Almost essentially, right?

(35:41):

Essentially. And we went out at that point and so that’s why we went down the US route. We got in contact with Jason Calas. We were approached to apply for the launch accelerated programme that he had recently launched Covid hit. So we’re talking three years ago. And so the programme we had applied for was to then we physically had to move to San Francisco and so we said, yep, we’re committed. Let’s do it. We put the application form in, went through the process, went through, I don’t know how many interviews, was it Fairview? Yeah, you

(36:15):

Have

(36:16):

Four or five interviews to get through that process to the point of then talking to our customers who were using. So

(36:23):

Interviews as in you were pitching different people essentially.

(36:26):

We were pitching the launch accelerated team and the different staff that had involved, including Jason himself, to get through the rounds to then say, I can’t remember the numbers now, but say 2000 people applied for this launch accelerator. They only choose seven,

(36:39):

Right?

(36:40):

We made the top seven. We made the top seven through that journey. But that opened up our world to venture capital, raising, raising capital, what a tech startup really was all about. It was completely different from running an MSP to going to the tech startup world. And I’ve written

(36:57):

Starting a company in the us they required that you had to be a US company, all that sort of fun stuff doing that from a startup.

(37:03):

What’s Delaware

(37:03):

Or something? Delaware Corp. Delaware. So yeah, so this opened our eyes to the entire world. But the cool thing was, I guess timing’s the thing is that because covid hit at first they were thinking about closing the accelerated programme down and then I think within 24 hours they said, Nope, we’re going fully virtual. And so we’ll have full the first virtual cohort from a capital raising standpoint,

(37:29):

Beauty Capital Australia and did it or

(37:31):

Stayed. We didn’t leave because we didn’t have leave,

(37:33):

We left. You’ve even

(37:34):

Stuck there. We were stuck in Brisbane. So we did the whole thing virtual. We were up, were in the office in the valley at 1:00 AM, 2:00 AM 3:00 AM

(37:43):

Still did it to us. Time zone,

(37:45):

US time zone, yeah. They didn’t change to us. So we just did stupidly long hours for months on end pitching to 1,015 hundred angel investors all around the world to raise our million dollars that we raised to kickstart our Tribu journey. That’s the short version.

(38:05):

Yeah, no doubt this would’ve been lot of lessons learned through that process. Lots of lessons learned, no doubt.

(38:12):

Amazing people we met, we got to meet all the majors you’d think of, but then also the person who runs the innovation hub at Disney, all the innovation guys at Nike. We got to meet some really cool people. And I say all the majors, all the major VCs that are in the us, we got to meet.

(38:32):

Has that angel world changed Mutts? Remember? I’ve

(38:36):

Very tightened up. So

(38:38):

Was you invested in Uber and was another

(38:42):

Airbnb?

(38:43):

Airbnb one of was it? And he wrote

(38:45):

That book, calm was the other big one. Calm,

(38:47):

Yeah. Oh, was he even calm? Yeah. Okay.

(38:49):

He wrote the book Angel.

(38:51):

Yeah, which is all about that angel investing. So I want to say that’s almost 10 years ago that

(38:56):

It’s a while back. He’s written another one. I dunno if it’s come out yet. He keeps talking about that. He’s been spending time writing a second book.

(39:02):

So you’re still in touch with him.

(39:04):

We’re part of a few of the, yeah, because once you’ve got into the launch accelerator, you kind of become part of that community and you remain in there. So

(39:12):

Has it shrunk? Is that,

(39:14):

So two things. One about the community. There was, I think we were like the 200th co-founder part of the Slack community and the other’s 580. Oh

(39:22):

Wow.

(39:22):

Okay. So whenever we raise the capital three years ago to now, he’s still actively invested. Even though the market’s completely bottomed out, he’s still actively investing into somewhat 10 startups a month sometimes like that. So with the marketing bottoming out the last 12 to 18 months, the whole landscape’s changed. This is the mentality, not just Australia, but this is global. Now, as a company with an idea or a founder or an entrepreneur with an idea, there’s really, unless it’s the people, extremely well friends and family, you’re not raising capital or you’ve got track record. So say you’ve already raised capital 2, 3, 4 times from VCs and you’ve raised tens of millions of dollars, then having an idea, they won’t back you. So what you need now is you need an idea. You need product to market fit product to market fit is absolutely crucial. So what that means is you have not just a few paying customers, but tens or hundreds of paying customers.

(40:15):

Hundreds is a product market fit,

(40:17):

Product market fit. But the number one crucial, I guess, aspect of any technology founder now is recurring revenue. $100,000 MRR is now the number you need to hit before you go out to raise capital

(40:30):

MRR $100,000 monthly recurring revenue before you look at raising capital with VCs in this market, it’ll change again another 18 months, two years, it’ll change again. But talking to a lot of VCs around the world, that’s like, cool, you’ve got your product out there, you’ve got some traction. Come back to us when you’ve got a 70 k. Come back to us when you’ve got an 80 k. Come back to us when you’ve got a hundred K, then we can talk about raising capital. So I’m like, once you get to a hundred KMRR, do you really need a raise? It’s not really a

(40:59):

Startup anymore,

(41:00):

Right? It’s not kind like a business now. And yeah, it’s a big example. The only reason you in to raise capital then is to fuel the fire to hire the 20 BDMs and SDRs and just go out there from a sales and marketing point of view.

(41:14):

Yeah, interesting journey. I’m conscious of time, we’ve gone a little bit over, but what’s next for Lucas and Dean and Tribu and you guys are working on

(41:23):

Tribute these days is really focused on the vendor. So we’re doing vendor integration, so selling AI software to MSPs hasn’t really gone to plan. I knew building the product was going to be hard or actually I didn’t even do any of it. That’s all Dean. But I knew building an AI product wasn’t going to be easy. I didn’t think selling AI to IT, support MSPs was going to be difficult. We were an MSP and every MSPI spoke to also wanted the software, but it’s been one of the hardest sales journeys I’ve ever had to go

(41:53):

Through. What was the main objection and reason for No there? Accuracy. I mean it’s just saving time and everyone wants to save time and therefore be more efficient, make more money, right? So what was the main objection you were getting

(42:04):

The two main objections? Yeah. Why do we need to change? Everything is running as it should be. All the resources we have in place, we’re making money. Customers are happy, so they say everything’s fine. So why do we need to change? So change is a big thing. The number one factor though, this was brought up earlier, is accuracy, but our staff can do it better.

(42:28):

I would say even though its accuracy was part of it, it wasn’t so much that the AI would’ve been wrong, it would’ve been wrong for some of them because they all realised that actually we have not been doing a good job with our ticket data. So far we have not had good process or policy in place already. Many MSPs out there are quite haphazard, let’s call it. Not all of them. The ones that have good structure, they were able to implement it straight out the gate and it’s work to treat. But what we found is it kind of put a spotlight on the fact that they haven’t been doing things very well for a long time. Because of that, we would train and then the model, the AI get things wrong because well, they’d been getting things wrong themselves. So it does learn mistakes and I brought up the accuracy thing before, but essentially if you’ve got too many mistakes in there, it will start to replicate that. It does have some tolerance for some mistakes, but not a lot.

(43:25):

Yeah, I guess that’s almost the reason for a lot of people, not in general AI or automation in their business is that accuracy piece.

(43:31):

Well, that’s right. Yeah.

(43:32):

Your larger MSPs, they have a good amount of discipline typically, and they can pick this up, your smaller IT providers and things like that, they normally start one, two people, one man band, that normal terminology, they don’t have much in the way of process discipline, let’s call it. So they’re just haphazard, they’re busy, they’re going crazy, and as a result, as team members come on and join them on that journey, they’re equally as chaotic. So then you end up with chaotic data and that’s pretty hard to train.

(44:05):

The majority of the MSPs using our software today are the 30 50 staff plus the larger MSPs, and that’s only 5% of the market and 95% of the MSP market are the one to 30 staff members

(44:16):

Where you thought it would be more traction with the smaller market.

(44:20):

Yeah, we thought they’d be wanting to bring themselves up to a higher standard when, and they probably do want to bring themselves up to a higher standard, but they realise that they have not done things particularly well

(44:30):

In the past. Fair enough.

(44:33):

So vendors is our focus now, which means we want to get our software out to MSPs. How we do that, we’ve now integrating into a few vendors in the MSP space. And so we will most likely become that whole white labelled. MSPs won’t even know they’re using our software because Cisco goes through the vendor. We help vendors embrace and adopt AI because they’re all talking about it. So if we can help our vendors adopt AI just happens, we built the model for MSPs, then it’s a win-win.

(45:01):

Got us of time. So we’ll wrap up there, but thanks. It was really good to get some of your insight around ai, around what’s going to happen next year and how businesses leveraging it now and how your view on AI and what’s happening. So it’s an exciting time and appreciate you coming and sharing the insights.

(45:15):

Thank you. Thank you very much. Thanks for having

(45:17):

Us.