Jovan Strika of ElastaLink explains how AI agents work like a new hire: your data, your rules, clear limits, and a human who approves. For small business owners and solopreneurs.
GROWTH PILLAR: AI & Automation
WHO THIS IS FOR: SMB owners / Solopreneurs / Corporate escapees / Leaders building systems
WHAT THEY'LL GAIN: A clear picture of how AI agents differ from chatbots. Why data location matters. How to set limits and approvals. How to test it through the AI pilot.
Most owners think AI automation means a smarter chatbot. It doesn't. Until that changes, you'll keep re-explaining your business to every tool you buy.
Jovan Strika is the founder and Chief Vision Officer of ElastaLink, a technology company he has run for more than 25 years. In this conversation, he shows how AI automation works when AI agents act like a new team member. They get your business data, your rules, and clear limits. He explains AI workflow automation in plain terms and why it beats a plain chatbot. He also covers AI security, including why your data should stay in a private cloud, and why a human stays in charge of every approval. Jovan shares how scattered tools hold back process improvement, and how the @Hive pilot lets owners test AI tools for small business. This episode is for small business owners and solopreneurs who want to save time without giving up privacy or control.
Key topics covered
Guest links
Supporting persons mentioned: none.
— PARTNERS ON THIS EPISODE —
Next Steps
Bernie (00:22)
Question to the audience: What if your business software actually understood your business? Instead of making you explain it
Over and over again every single time. Our guest today didn't just start his company. He built and exited more than one company before it. He built this one from the ground up, and 25 years later, he's still the one holding the pen on where it goes next. His official title and all, Chief Vision Officer. So that feeling of explaining yourself to your own systems is one of the most things that business owners know too well. Today we're going to
Talk with Jovan Strika He is the founder and chief vision officer of Elastalink, a company he's built for over more than two decades into a technology partner for businesses, governments, and growing organizations across Canada. Jovan studied usability at Concordia University and e commerce in London. And that mix shows up in his work. He cares as much about how something feels to use as how well it works. Before Elastalink, he built and sold other companies.
So he knows what it's like to start from nothing. These days he's focused on something bigger, helping businesses use AI without losing the human judgment. That makes a business worth running. He splits his time between Montreal and Ottawa, where his team is piloting that idea through a project called At Hive.
You can find them at elastalink.com, spelt E-L-A-S-T-A-L-I-N-K dot com or on LinkedIn under Jovan Strika J-O-V-A-N-S-T-R-I-K-A. Jovan, absolute pleasure having you here today. Thank you so much.
Jovan (02:03)
Bernie, pleasure is mine as well. And thank you so much for having me. I was looking forward for our session.
Bernie (02:10)
So what were you doing just before you started Elastalink
Jovan (02:14)
Well it's interesting my journey. I think if I go back to my beginnings probably will need another dedicated episode for this, but would say that starting from Europe, where in Yugoslavia at the time, which kind of split into multiple countries, and then end up being in in Spain, Cyprus, UK, then moved to New York.
and that's where I realized in nineteen ninety nine the voice over IP is taking over telecommunication that prompted me to decide to move to Canada where I started my first company here and that's been as I said, more than twenty six, twenty seven years.
Bernie (03:00)
Has there been any major lesson that you learned from any of your previous companies that you still use today?
Jovan (03:06)
Yeah, so interestingly when in my beginnings when we talked to the investors, when we talked to customers, it was always about being focused with what you offer and dedicate your company around one solution at a time. and that still holds the truth for many businesses out there. but what we started realizing from previous companies that once you establish trust and when you
are able to introduce and cover multiple services, customers really appreciate that. Especially when they're not technology, they like kind of a concept of one, one stop shop or one one phone call to make or one throat to choke, depending on how you present. And that's the value that I said resonated and confirmed many times over and over from company to another. and that's that's what I'm still believe that things change now.
that you you need to offer complementary services because then you you have ability to provide a different level than being isolated island. so that's the lesson I would say that it and mantra that I'm still using.
Bernie (04:23)
So you you've started Elastalink is your primary company. You've started another company where Elastalink helps support it. what's that all about, At Hive?
Jovan (04:35)
Yeah, so it's interesting. When you look at Elastinink as a company that covers all major services that businesses would need from IT and telecom. as I mentioned, instead of dealing with seven, eight different providers, Elastalink was covering all the IT and telecom services. But what happened, especially during COVID, I realized that the future work is different, things are changing rapidly. And
We kind of brought from the point that technology is critical for all business sectors. we start looking into At Hive as a concept where very similar idea resonates, but on a business level, not just on a technology. definitely technology is a driving factor in in At Hive, but we're looking from the space, building community of business professionals, and helping them with
using technology, how they can increase productivity and really transform themselves to accommodate better future work.
Bernie (05:40)
So one of the things I've discovered and you've already you've already started working on a solution map is that there are different services that we all need, whether you're a solopreneur or a one to five man team or larger, even especially more for the larger groups. 'cause you have your Elastalink team, you have your At Hive team, you have the at Bistro, so you have
Provide catering and and food services, all that is coming in. what's the actual cost to a business if you don't pay attention to these things?
Jovan (06:12)
Yeah, this is the a challenge where th your cost is fragmented between many different providers and different services. then really bottom line is something not visible immediately but only later. And the challenge obviously is that it's not just about what hits your pocket, it's about also how much of that productivity you lost. are we able to advance? There is a different
values that might not be quantified in just a dollar sign and with a dollar sign, but also how you adjust yourself towards the the new changes that are happening. we're seeing uncertainty on many different factors from impact on AI and potentially robots, to wars to trades to so many different
impacts that are happening right now and then how you quantify effect of that moving forward because it's not obvious immediately it's a long term change. and this is where At Hive was positioning as a way to to really accommodate that change and help businesses to first educate and point but also to connect people to create a community that can help kind of us together
come come over that trend and being able to adapt to the changes rapidly because doing alone it it is very difficult. and that AI certainly we saw the impact as one of the biggest changes that is happening worldwide. internet was one of disruptor. AI probably is the bigger one than than internet. And this is what I'm
hoping that with what we're doing here it's just the beginning that will help businesses to adjust because I believe it is window and the companies who are not able to adjust quickly enough will unfortunately pay a big toll down the road.
Bernie (08:22)
One of the things I noticed, and you've done this with some of the other services that you offer through ElastoLink, and now you're doing it with the AI component, you're bundling it all under one house. if right now if I set up my office, there's the word processor, there's the email, right? And then when I do my video editing, I've got little bits of everything all over the map. You put this all under one house.
Is that the other part of the efficiency to allow for scale and versatility and inter opera interoperational capacity?
Jovan (08:55)
Yeah, it's it's a great question. when you look at just from the fact that when cloud start really introducing solutions for the businesses, you have cloud applications that are fixing one one thing or providing you solution for one thing at a time. But at the same time, it is the isolated island that the data and that application sits in. so the work in within that application
That is happening, the data related to this is isolated from the rest of them. So there are some integrations between them, but in most of the cases the context that is related to the data that is stored does not travel between. This is the premise where we saw opportunity with the Lastalink, where we as a provider of telecom services are introducing the main
applications for communication. So phone, video, messaging being a corporate chat, files, the documents as well as the email. And from the high level, when you look at the understanding what's happening within the organization and people who are working, with access to the data that is exchanged internally or externally with the customers or investors,
pains the a really better understanding of what the business is about and therefore with that context ai could help putting in solution in place. Otherwise businesses would need to grab the data from different places and put this in the context that AI can understand. And that takes time and more burden to the professionals and businesses to do this and invest the time.
So having access to data is number one thing that AI then can really be empowered to better to provide us services and to better understand the business and therefore being able to understand and and learn about a business continuously improve and adjust. But what we are looking here from the point that AI is not self
self-driven. It's a critical that it is the human direction on top of it that stays in the center. That people are defining processes, but AI can support them, improve them. approval still needs to come from a human and then having a very clear documentation and evidence of all the all the changes and all the implementations that are done. And this is the I would say
what we see in lately AI and especially now the next generation, because we are think in just only six months ago, we were when we talk about an AI, people were really thinking more about a chat and I ask a question, I get an answer. We're evolving into the the next level of AI, which is this agents or agentic agentic frameworks that are
providing a very different level. and that's where we're seeing originally what was called a large language model, which is the when you have a chat and you get an answer, is now elevated through taking these capabilities, but together with the context of your data and documents, conversations, business records, even your databases that might be custom.
being able to allow AI agents to access this data, then together with the what you call skills, companies being able to capture their processes and explanation of how they they use this data, how they perform the businesses. not not very different than hiring somebody else that a new employee that you need to teach them and give them access to the data in order for them to do the work in a very similar way.
You need to treat the A agent with the same capabilities and then expose them with the tools that they can search and access this, create or update and on top of it humans providing observation, assigning permissions that they cannot do what you're not allowed to do, is then
the I would say the letter and and stack that really changed the look from AI agents almost as a a business employee. and in context of exactly what is exposed, you're now s seeing it as a virtual assistant that either can help your current employees and or be self autonomous that can be just contributing to the team in order to
progress and bring that knowledge for the rest of teams to to make decisions. So it is a very different, I would say, approach, which we start seeing about a year ago, how that can impact the businesses. And that's what we've been lately working on and seeing how within the At Hive, where the business professionals are having an access
we could give opportunity for different members in their businesses to test this concept and start improving their workflows.
Bernie (14:49)
I like the part that you still have the because I hear this a lot, human has to still be part of the the solution. And I'm I'm gauging also that here I use a second brain. it almost sounds like the scarecrow moment, but you you already have the second brain built in because it has the context of the entire space all in in the one platform, which is also really handy. That way you don't have to keep, searching out information.
also from our past conversations, I also understand everything that the large language model is working on, everything that it's doing, is behind the four walls. It's not out in the cloud somewhere else, in the the public cloud. And it's all here in Canada. is those are kind of key factors, are they not?
Jovan (15:40)
Absolutely. I mean this is one thing that with our experience working with the businesses, we realize that a lot of them don't understand the implication of using AI when especially when it comes to the privacy. every question that they are sending to Chat GPT or different AI tools is executing this within the public cloud, which means that your data and it can be very sensitive data.
is potentially sent to another providers in other countries in in most of the cases. and eventually if it's not well easily detried it can end up being used as a data set for training. and this is a challenge that we're seeing that just the the bit the businesses are not really understanding implication. and
Some of them are making decisions who completely ban the use of AI. even if that's detrimental for for their business, they just don't see the way to to really isolate that. our approach is different from the point that we are Canada based. we're local, we have our infrastructure that we created because of the other telecom and IT services. And we leverage that to actually
process everything including hosting of applications as well as the data locally. that is what we call the private cloud which is sort of sits in between your private data data room in your office and the public cloud meaning that information is stored locally or close to your business depending where you're located.
But it is not shared with anybody else. it's sitting on a dedicated and isolated technology that only you as a business have an access. And that is the how we formalize as a private cloud. we can then ensure the privacy of your data. And that insulation is very important for a lot of businesses. we tend to not use what are called the multi-tenant systems where
resources are shared. but it is also question about how who is managing this, who is taking care of the backups monitoring, which is the part of our team. so it is a bit of unique proposition and more importantly, still be affordable for our small immediate size businesses, which is unfortunately the challenge with a lot of this new advanced AI applications and AI agents is that
Microsoft and other larger organizations are primarily able to focus on the the government and the large businesses, but the the small and medium sized businesses still are are having the same type of problems and challenge and it's very hard to fit in their budget, which is what we've been trying from exception of the last talent to to really bring an enterprise grade solutions to be affordable by small and medium sized businesses.
Bernie (19:00)
I think one of the other aspects is that let's say if someone needs a phone system, they can get a phone system. If they need other services, they can buy it in components as opposed to no, you have to buy the entire suite. so yeah, you have that also that leverage that you can work with. So and
If you have that chat bot inside your LLM it it continuously learns. It's almost yeah, at a certain point the more you use it, the more that goes through it, the the the smarter the system becomes almost like a little Yoda.
Jovan (19:33)
Yeah, great analogy. So you're absolutely right. I mean I wish to say that I have a foresight ten years ago where things will go. I would say that without a lot of planning we end up in a very unique situation where we as a provider of business communication and and IT have access to
to the data that to customers if they want us obviously to leverage some of this technology capture as an input being able then to connect them together all this independent conversation including your video call today with a zoom you can get a transcript of your meeting but is there something that is connecting all of these different meetings?
that is happening across your organizations. So think about how if something like this can be connected in context of customers and your team, you start paying a painting a very different picture. And then being able then to find relative information, either human or agents, to extract the thread and then apply and use in creation of new context analysis reports.
this would not be possible if Velaston did not really start with being kind of a one stop shop when it comes to the IT and telecom. Now another aspect that we did not mention but equally important is cybersecurity. And we're seeing AI used for by hackers for for finding the ways to penetrate organizations. and cybersecurity unfortunately
is also applied often in an isolated way. whoever is a provider of solution, they address the cybersecurity, but really should be looked into overall as an umbrella between all your services and not treated independently. so when we look in from a cybersecurity perspective, we were looking for protection across all the applications. and
I would say between AI and and cybersecurity, that gave us really ability to proposition this type of service in a very different way than than and other businesses.
Bernie (22:05)
So as someone's looking at this and going, Ooh, I I would wouldn't mind having some of that. first off, this is all in Canadian pricing, and can they get quotes on certain components then as opposed to just, a normal little chat bot that you would get for twenty dollars a month from pick one of the major LLMs that's not secure and tight and not integrated yet, or you have to learn how to integrate everything, as as I'm finding out.
APIs and other things and other, or there's other instructions you have to put in there. So it yours is coming self contained, s built ready and just put in your profile. Is that correct?
Jovan (22:48)
Yeah, absolutely. I mean it's it is the journey, it's the decision that starts with a small steps. we need to walk be f before we run. and the idea was that through this journey that the businesses and professionals are able to again with each step value and see how that can transform the way they work.
we were as you mentioned about is the first step having a more secure chat. Absolutely. I would say maybe the step before was part of education and we did launch a year ago AI consultancy with a purpose to to really have a dialogue with organizations to understand where is the opportunity, how AI can help them.
because they're hearing all of this AI can do the great things they personally experience with chat GPT the value, but they have a hard time understanding how that can impact their business. Because it is different. It is customization, pro-organization, or even from an individual perspective, from your title and your department and and and what kind of job you do. So the Air Consultancy was a lot about first
understanding areas where we can without a lot of effort already provide a value. and then moving towards predefining some of the what are called skills, so we call it experts, teaming up with the different areas, different departments with the capability, and then slowly looking into also other type of information. So
when you look at a business application, CRM, inventory, e-commerce, all of these are customized applications, depending on type of the business, where also data is, how much of that can be linked together or automated to speed up the process? Idea was to take a lot of that I don't want to call it boring, but kind of repetitive tasks are ideal for turning that into automated process.
and giving that time back to employees, that then they can do and spend the time for other areas that AI might not be able to help. So how you turn that into from personal assistants or team agents working together with the with the rest of the of the work but also learning as they go. I want to highlight one thing this that AI agents and whatever we set up
initially is the starting point, but it's not endpoint. It is constantly evolving through all the new data and experiencing and fine-tuning over and over. So it is a journey. it's not kind of a let's start with the signing up and activate a service now type of thing. it is the decision, but certainly we see return of that investment in multiples of that return
once businesses start using it.
Bernie (26:09)
So when someone starts using the platform, where where should they expect the human operator to be part of that process? Do they is it when it's done its final checking looking for certain key points? is there a a dashboard, anything that kind of pops up, or is there something that says, hi, I got to this point, are you happy with it?
Jovan (26:32)
Yeah, so it's a it's a great question. and I I maybe want to spend a little bit of time explaining that kind of evolution of that interface. when you when you look at communication stack emails, phone calls, chats, is how humans are communicating. The very first introduction of communication with the AI
really I would say mass marketed or mass adopted interface is the chat. that's how even with AI and machine learning that was happening many years, people did not really understand that until Chat GPT was introduced on the market and end up in a very short period of time getting 100 million users. That first
Interface is the the chat system where you ask a question, you get a bubble back with the answers. And although it is revolutionary because it is using natural language, you communicate, chat is not a really great interface. and that started involving into now being able to use the voice. also the messages are not maybe necessarily text, they're attachments and tables. But
What I want to bring up is that we're on the verge of evolving that interface into something what we call interactive workspace. so instead of actually having a just a plain text, what we are start building is this interactive snippets of the user interface that you can interact with. So if you ask for for example
for a specific quote, you can get a sort of a preview that allows you to click a buttons, maybe sort, search, type something. It's kind of hybrid interface where you can mix and start with the text but then switch into the traditional clickable interface. But really only for the snippets or portion of the interface that is relevant to your questions, rather than going through all the pages and
interacting with the website and the navigation. This is sort of in place information that now you can you can get in the chat window that will allow you to click through and maybe open something else from there. and we noticed this idea supported by Google and other
researchers. matter of fact they even introduced the the protocol called A2UI. idea was the agent two user interface as a way to provide a better interactive interactive method. we kind of brought up to to be more declarative with the some of our application framework and being able to adapt
sort of real time between the type of information that you are requesting, because obviously that can be very different from the business to the business. but it is clearly on a on a on a beginning of that change that I'm sure that you will be experiencing shortly.
So
Bernie (30:15)
Someone starts capitalizing on the AI. Are there any surprises, anything that they might not want to see? or is that pretty it's pr pretty safe. It won't commit you to yes, we'll buy a thousand widgets and we only need two.
Jovan (30:31)
Yeah, so it's it's a really interesting when you dive deeper into what is required when you try to put this type of power in a context of the business, you realize there is a many other challenges that we had to overcome. what we are seeing most of the AI experiences today are on a personal level. including
some of the recent AI agents experiencing from open claw harness agent and others there's a coding agent that are helping developers to create a new software but all of them is running on your local computer and you are allowing agents to use your local applications and and that open
a lot of security questions. What if they send an email or spend your spend your sp spend your money on purchases or just open a door or install something that you did not want to allow or you were not aware that that can compromise your system. we had to approach this in a very different way. We had to first create a restrictions between what agents can do by very clearly defining
what type of tools or what type of applications agents will have an access. And then there's a second layer of that protection which is based on employee. So if you are asking agents about maybe your vacation policy, that can access only your specific contract and your own
information that only you can see, where if I'm asking them I get a completely different answers. But if I actually access certain customers or ask for seeing invoices, if I'm not in the finance department, I may not have at all access to this type of data. So the big part of what was tricky is to create this first what agents can do and constraints. And secondly is what kind of access they can have. And typically when
we have an agent that is assisting you as an employee, we are then passing or inheriting that permissions what you as employee can access to the agent and in that way restrict. so that is I would say probably one of the toughest part that we're trying to to implement really bringing this into in a secure way to the public.
Bernie (33:24)
it begs another question though. Let's say you establish the the platform inside a business and it's running. And can you as the owner or the administrator then look at okay, where should I go next? am I missing if you're looking for gaps, right? Some things you haven't paid attention to in the past. Can it help find that?
And at the same time, almost like self-healing, can it go, hey, you should add this service. So is there kind of a loop that comes in and says, review this element? whether it's let's say it's doing a search, right? And you're telling it to go find, I gotta pick a restaurant, I gotta take the crew out for Christmas dinner or whatever the case might be, and
Allow it to go hunt that information out and find best pricing or, something that fits the food profile, for example. I mean Yeah, sorry, go ahead.
Jovan (34:18)
Yeah, absolutely. I mean i it is again I always say think about as having just another human being that is assisting you. and it's about what you expect from that that human being to do. When you ask for a task, will you tell them just make the best decision and just let me know when it's done? Or you will
ask, gather context, draft certain things, present to me so I can approve it and then go and execute. and the good part of this is that self learning. So for certain tasks over time it might understand that, I can be autonomous and I can handle stuff. and for some not. Like a good example is the agent helping you just with your email.
being able to read your messages, understand that some specials and promo emails are not important that might delete or leave it for later for you to review. the ones that are a more informative email saying, this is completed and they just CC you it can gather all of this information from multiple emails.
and prepare for your morning update to tell you in a one glance what what happened so you are informed. And then for the ones that are requiring your attention, it can draft the email content and you can decide if you want to change it or you're gonna just execute and send. And this is the different levels of how you really fine tune and what you expect from your assistant.
and you can change it as you go. So this is the level of I would say human driven agents that it's very important to adapt to your preferences. Clearly there are some checks in place that if it's something done, for example, a certain change, there is an audit log that is very important from compliance perspective.
being able to see what was done and that trail record has to be in place. and certain transactions certainly can be revert. but agent can put things in in the basket and then let you to review it and decide to press button, pay now and place the order, and that's up to you.
Do you trust agents to do this and with your wallet or not? it is something that can be configured.
Bernie (37:02)
So tell us more about At Hive and what you're hoping businesses can do with it.
Jovan (37:07)
So At Hive is it's a really a great platform that independently from what we're presenting here with the agent, I believe it's ecosystem that was missing. and what we're doing is as you might mention in the beginning, it's not about a co working space. It's not about just a place to work, although
Anybody who is in Ottawa, I would welcome them to come and visit Arhive and experience sort of future work and how we design the the space around zones and activities that fit what you do at a time rather than having all of a single desk from where your work is expected to be done. but we we look into other aspects of the Arhive, really building community and there is a six
areas that we we see as a combination of all are what is impacting business sort of future or work. technology is one of them which I kind of explain quite a bit, but the learn is another big pillar and learn not only from the context of personal learning but also from the context of your knowledge transfer and
We saw that as a big challenge, especially during COVID businesses trying to hire somebody to the organization had no really a a great way to transfer that knowledge. we we feel that that is the first step in creating these business processes. Some of that can be learned through observation, like what we talk about in agents having an access to the data, but a lot of them is something that
organization can document now with the help of agents and AI much faster and easier. the other one is the connect which is really creating their community and allowing other members with different expertise to help each other. And especially when we look at local businesses and local community, we feel that that's very hard to do lost over time. We're trying to really bring that focus on
on local community and helping each other. the balance is very important part as well, which is when it comes to Bistro the food always is the great way to connect people. and welcome anybody to come and join our Bistro as well. in general all of these are what I'm explaining pieces that are really influencing and impacting
from different angles how business can be successful. And we're trying to with Ottawa launch introduce that concept with an intent to really grow and introduce that in other areas and even beyond Canada. because there's a lot of similarity we see across on a globally a need for a different approach. AI and applications that we offer as a part of it.
you don't have to be physically locally in Ottawa. you can have an access to a lot of that just being online. so don't don't feel bad if some of our listeners are not in Ottawa. We can still help you out and grow your business.
So I just want to mention At Hive is the first public step. we felt that it's a good platform to give an access and to engage with the first organizations. and from there we can start customizing further. so what I would suggest to follow up this channel and
will provide a links to be able to schedule one on one sessions, enroll in some of the the work we're doing and have access to first agents that we're launching under the At Hive platform.
Bernie (41:24)
So folks, if you've listened to this podcast and you're going, hang on, I'd like to know more. I want to learn more. Jovan's team is opening up their At Hive pilot to businesses who want to test drive this. And you can put your name in at At Hive, so at hives.com, and just look for join the pilot.
Jovan (41:45)
it is certainly opportunity to start a dialogue and to by inviting everybody from a business perspective to start a journey. this is beginning of the change that I think it's imminent and I I feel really important for businesses to react and looking forward to engage with some of you.
Bernie (42:10)
Perfect. And I I think there are also incentive programs too for certain businesses. Are there not?
Jovan (42:16)
Yeah, so Canadian government introduced several different programs that if you are incorporating Canada you can take advantage. we're we're seeing other countries are doing a similar concept. that that's part of what we can introduce once we start a dialogue too, depending on where you are in your journey, is established business, is a new idea that you're trying to launch.
different incentives are available. so we can definitely address that as a part of pilot
Bernie (42:51)
That's great. So folks, Jovan's living his results and he's building them so that you can enjoy them too. And having tools scattered all over the place and it this unifies them. It's under one roof and it's easier so that if your phone's hooked up to your AI, it it just integrates so much cleaner, right? And and then you have chatbots and other supports and it's it gives you a workspace you can
build on and and grow and scale your business with so building it up and taking the tension out of the day-to-day stuff and taking away the what's gonna happen next moment because at the end of the day Elastalink is there to help you and grow with your company and if you want to reach us we're at knack4business.com that's K N A C K the number four business.com
We have a chat bot on their bottom right hand corner of the website and you can inquire about other podcasts or book a call. Jovan I want to say thank you so much, sir.
Jovan (43:53)
Bernie, so thank you very much for having me. it was a pleasure and I'm looking forward for the future episodes where we reveal some new stuff as well that your audience can benefit from. Thank you.