In this episode of Software Spotlight, host Michael Bernzweig sits down with Steve Flavell, co-founder and co-CEO of LoopUp, to explore the company's transformative journey. Steve discusses how LoopUp pivoted from a conference call software company to a global telecommunications provider, adapting to industry disruptions and the pandemic's impact. The conversation delves into the restructuring of the telecom industry, the move towards cloud-based solutions, and the integration with platforms like Microsoft Teams. Recorded live at the Software Oasis AI Summit, this episode offers insights into resilience and strategic adaptation in the face of change.
Introduction
- Steve Flavell's 20-year journey with LoopUp
- The shift from conference calls to global telecom
- Surviving the pandemic's impact on business
Interview
- Founding LoopUp with co-CEO Michael Hughes
- Initial challenges and opportunities in conference calling
- Navigating industry changes and pivoting the business
Presentation
- Restructuring of the telecom industry
- PSTN switch-off and its implications
- Integration of telecom services with platforms like Teams
Panel Q&A
- Governance checkpoints for telecom restructuring
- Strategic adaptation in highly regulated regions
- Audience questions on AI integration and telecom strategies
About Steve Flavell
Steve Flavell is the co-founder and co-CEO of LoopUp, a company that has transitioned from conference call software to a global telecommunications provider. With a background in engineering and strategic consulting, Steve has navigated LoopUp through significant industry shifts and challenges.
[MICHAEL]: Imagine building a company that rides the waves of technological change for over two decades. That's exactly what Steve Flavell did with LoopUp. Starting from humble beginnings in conference call software, Steve and his partner navigated the shift to global telecommunications, all while surviving the pandemic's impact on their business. And by the end of this episode, you'll understand why Steve's story is a testament to the power of reinvention and strategic pivoting. I am your host Michael Bernzweig and this is Software Spotlight.
--- INTERVIEW ---
[MICHAEL]: for a lot of the individuals listening in today, they've all had their own journey, getting to where they are.
[MICHAEL]: And for many, that journey has been a journey filled with a lot of side trips and potholes along the way to success.
[STEVE]: So I was hoping maybe you could share a little bit of your journey prior to LoopUp.
[STEVE]: And how you got to where you are today.
[STEVE]: Of course, of course.
[STEVE]: I think the LoopUp that you see today, and we've gone into this a little bit, which is the global telco, trying to consolidate how multinationals buy and manage their telecoms around the world, or all integrated with teams.
[STEVE]: But this has only been going on for five or six years, this ability to integrate your telco into teams.
[STEVE]: And LoopUp's been going on for 20 years.
[STEVE]: So as you can probably tell, we've had some journey along the way before getting to the bit that we talk about with, I hope, some passion today.
[STEVE]: And that background for us, that journey for us, started actually in the world of conference calling.
[STEVE]: For 15 odd years, we were a conference calling or remote meeting software company, not a global telco.
[STEVE]: And we had a product that was all about making conference calling less bad.
[STEVE]: It's changed totally today, particularly since the pandemic.
[STEVE]: But if you think back to pre-pandemic days, when you were dialing into conference calls with numbers and codes, and you had no idea who was on or who was speaking or who had all the background noise, and we entered LoopUp in 2006 to make that experience, that important experience for companies, less bad.
[STEVE]: And for a number of years, it did really well.
[STEVE]: We grew consistently for a dozen years.
[STEVE]: And we managed to get ourselves public on the London Stock Exchange.
[STEVE]: We built a close to half a billion dollar enterprise value market cap company.
[STEVE]: And the pandemic didn't kill us, but it damaged us badly.
[STEVE]: And we had to reinvent ourselves in the pandemic, or coming out of the pandemic, when those conference calling services moved from making that experience less bad that I just spoke about to a totally different category, this unified communications category that where meetings are all on video.
[STEVE]: It's a feature set within a much broader category known as unified communications.
[STEVE]: And that damaged our business badly.
[STEVE]: It set it into decline.
[STEVE]: And we had to reinvent ourselves with a brand new service and come back from that challenge.
[STEVE]: So in a nutshell, that's the history.
[STEVE]: And what we do today is product number two.
[STEVE]: It was an almighty pivot after 15 years of being in business.
[STEVE]: And that's the reality of business.
[STEVE]: You're always reinventing yourself.
[STEVE]: I think they've often said what doesn't kill you makes you stronger if you survive and come out the other side.
[STEVE]: So it sounds like that was the case for where you're at today, really a stronger organization.
[MICHAEL]: So can you share a little bit with the listeners prior to founding or, I believe, co-founding LoopUp, if I saw correctly?
[STEVE]: Yeah, with my business partner, Michael Hughes.
[STEVE]: So we're 20 years in, and we're still co-CEOs together.
[STEVE]: So I know it doesn't work really well.
[STEVE]: We seem to have made it work.
[MICHAEL]: So prior to founding the organization between the two of you, what was life like before LoopUp?
[MICHAEL]: What were you doing?
[MICHAEL]: What was your journey getting to that starting line?
[MICHAEL]: Yeah.
[STEVE]: Well, we actually, perhaps not the advised thing for co-founders, but our resumes looked rather similar, actually.
[STEVE]: We came in both as engineers by training at a university, both went into strategic consulting, both went together, which is where we met, to Stanford Business School for a couple of years.
[STEVE]: And then coming out of that, both entered the startup world.
[STEVE]: And it was good that we had very different desires as to how we wanted to spend our time in LoopUp.
[STEVE]: Michael looks after all of the technology and clever side of things at LoopUp, and I look after all of the sort of commercial and financial and investment side of LoopUp.
[STEVE]: So we split ourselves that way.
[STEVE]: Fantastic.
[STEVE]: When we got into it, we became free at the same time.
[STEVE]: And we'd always said it would be good to do something together one day.
[STEVE]: And back in 2003, we were both available.
[MICHAEL]: And we thought, well, if not now, then when?
[MICHAEL]: It's a perfect time.
[STEVE]: And it wasn't that we had a burning business idea.
[STEVE]: We just thought it would be good to go into business together.
[STEVE]: And we sort of made a pact that we were going to build a company together by hook or by crook.
[STEVE]: And if we didn't get it right first time, we'd have another go, et cetera.
[STEVE]: So we debated a whole bunch of different business ideas for two or three months, and then finally shook hands.
[STEVE]: We were actually in a Las Vegas casino one night, and we shook hands on what became LoopUp.
[STEVE]: And about five minutes later, one of the guys we were playing blackjack with decided to invest.
[STEVE]: So that's sort of where it all came from.
[STEVE]: I love it.
[STEVE]: I love it.
[MICHAEL]: Now, what was the original challenge or opportunity that you saw that needed solvent in the marketplace at that time, because that is a very different environment?
[MICHAEL]: Yeah, so it was in the world of conference calling.
[STEVE]: And we actually built one of the first ever BlackBerry apps for business, actually.
[STEVE]: And the way that we went about trying to make that experience less frustrating, less annoying, there were some wonderful YouTube videos back then making humor out of conference calling.
[STEVE]: And we used the BlackBerry, essentially, as a remote controller.
[STEVE]: We built an app that effectively turned your BlackBerry into a remote controller of a conference call.
[MICHAEL]: And when I say that today, it sounds trivial, right?
[MICHAEL]: The things that iPhones can do and where they've moved to.
[MICHAEL]: But back then, this was sort of a pretty groundbreaking app.
[MICHAEL]: And when you showed it to people, it was almost like magic in that you had your BlackBerry, and you could see who was on your call, and you could see who had all the background noise and mute them.
[MICHAEL]: And it sold beautifully into the world of important meetings.
[STEVE]: We pitched it to the legal world and the corporate finance world, where those meetings are so important that they go well, and where you actually found a lot of BlackBerrys back in the day.
[STEVE]: Yeah, it's a blast from the past.
[STEVE]: Yeah, and it really took off.
[STEVE]: It was the first startup that I'd been involved with where it sold straight away.
[STEVE]: The go-to-market, the product market fit was there from launch.
[STEVE]: And it was almost there too quickly, because our launch version wasn't necessarily quite ready for prime time.
[STEVE]: And people wanted it.
[STEVE]: And we knew then we were onto something.
[STEVE]: Yeah, you were in the pocket of every business person.
[STEVE]: I mean, BlackBerry was it back in the day.
[STEVE]: So you mentioned app, and as really BlackBerry as the controller.
[MICHAEL]: Where did the app live at that time?
[MICHAEL]: Where was it?
[STEVE]: Because today, we look at Google as an app store, obviously Apple as an app store.
[MICHAEL]: But where was it at that time?
[STEVE]: There were special mechanisms to get, I forget the name of what they were now, it was so many years back.
[STEVE]: As an enterprise, you had a management layer where you were able to push applications down onto all the BlackBerrys of the people who worked for you.
[STEVE]: So our implementations weren't particularly going off to a store, although those did come along.
[STEVE]: And we did sit on stores within a pretty short space of time.
[STEVE]: But they were much more the enterprise that was buying your service pushed the app to all devices on their estate.
[STEVE]: And it was done in a managed way in that sense.
[STEVE]: Yeah, interesting.
[STEVE]: Well, if they could have only seen the future, maybe things may have transitioned differently, but any which way.
[MICHAEL]: So today, you've come full circle, but what was it like in those early years going from zero to one?
[MICHAEL]: How did you find those first customers?
[MICHAEL]: Was it BlackBerry that really launched where you are today?
[MICHAEL]: Or how did you bring aboard those first clients?
[STEVE]: The first clients were people we knew.
[STEVE]: Both Michael and I had been to business school.
[STEVE]: And a lot of our classmates were in the world we were selling into.
[STEVE]: They were VCs or bankers or lawyers.
[STEVE]: And we leveraged our network in the early days.
[STEVE]: Later on, we built a real go-to-market machine, a very scalable approach to getting this out to the market.
[STEVE]: We used a selling method that we came up with, which was called Pods.
[STEVE]: And Pods was all about team-based selling where you sort of tried to get away from the classic sort of sales issues of salespeople sort of hogging opportunities themselves, not sort of getting the right mix of team members in on an opportunity for fear of losing them.
[STEVE]: And we used team-based incentivization in a sales context to change all that.
[STEVE]: And the way that we found that worked well was to not hire people who had done any sales.
[STEVE]: We only hired people who had never done sales before because they weren't contaminated with the different approach.
[STEVE]: And we went all new grads.
[STEVE]: At one point, we had over 100 people in our sales organization, all but two of which we hired as new grads.
[STEVE]: So it was sort of our own methods for go-to-market.
[STEVE]: So coming full circle for a lot of listeners who were struggling with telecom and bringing it all together into a unified space, many have mobile infrastructure and traditional infrastructure and they're trying to bring it together.
[MICHAEL]: Can you talk a little bit about teams and that component of how you've really done something very different than what's out there?
[STEVE]: Sure, so this is our world today.
[STEVE]: Forget conference calling now.
[STEVE]: So this is back to what we do today, which is global service provision of telecoms.
[STEVE]: So the proposition is essentially vendor consolidation and simplification for relatively multinational organizations around the world.
[STEVE]: And, you know, that proposition is one of simplification and cost reduction.
[STEVE]: You know, the enterprises we sell to manage now their core, really their primary communications tool, which is nearly always Teams, Microsoft Teams, and they manage it in a singular global way.
[STEVE]: It's all done to one set of one, what's called one tenancy, one set of policies and procedures and methods.
[STEVE]: And why on earth would companies that now have that consistency of primary tool want a whole host of telcos, which is still critical, but perhaps secondary now to the time they spend on things like Teams.
[MICHAEL]: Why would they want a whole host of disparate telcos that have to be managed in different and individualized ways with different contracts and, you know, different support setups and tariffs and implementation methods and different features even?
[MICHAEL]: Why would you want that when you can have it all integrated into your primary globally run capability, all done in a sort of consistent, consolidated way?
[STEVE]: You know, we find, you know, it's a lovely proposition to go to market with because it doesn't take much understanding, you know, the core essence of the business case.
[STEVE]: Yeah, and, you know, at the end of the day, I mean, looking at the, you know, the number of enterprises using the organization, it absolutely, you know, speaks for itself in terms of the simplicity and the value proposition.
[MICHAEL]: Steve just shared how LoopUp evolved from a conference call company into a global telecom player. Now, he's going to dive into how the telecom industry itself is changing and why these shifts create new opportunities. It's a great continuation of how resilience and adaptation can drive success.
--- PRESENTATION ---
[STEVE]: Very good.
[STEVE]: Well, let's dive in and give it a go.
[STEVE]: I was keen to talk today about something that may sound surprising and try and convince you all of something that, you know, as well as all the very exciting AI stuff going on in the world, even some of our older industries can be exciting too.
[STEVE]: And I'm going to, in particular, talk about the world of telecoms, which is the world that LoopUp is in.
[STEVE]: It's an industry that's close to a couple of hundred years old, and it's kind of strange, but interesting how much is happening in that 200-year-old industry and leading to a lot of opportunity for companies in that space, despite its age.
[STEVE]: And that really stems from quite a rare combination of circumstances, pretty rare to find an established industry of such huge size, you know, just the, we're talking business telecoms here, the global market for business or enterprise telecoms stands at over a hundred billion dollars just in terms of service provision, no equipment or anything like that, just services that companies buy, phone service essentially.
[STEVE]: And at the same time, that huge industry is restructuring at such pace and on so many sort of layered dimensions.
[STEVE]: And some of those restructuring dimensions are further progressed than others, but all of them over the last 10 or so years, and many of them are brand new, and that pace of restructuring, that combination and layering of restructuring is really changing the dynamics of this established market and leading to opportunity.
[STEVE]: So let us walk through that a little bit, starting with the first dimension of restructuring and the one that's pretty much done, actually, this has been going on for a decade plus more and is all but done, which is the so-called PSTN switch off, it's how it's referred to in sort of the UK in particular and around Europe, more referred to as the retirement of copper in the States, how the PSTN, which is the public switch telephone network is disappearing.
[STEVE]: It's being turned off and replaced by voiceover internet protocol, VOIP networks.
[STEVE]: Essentially how the big telcos fit into this particular restructuring, they're the drivers, they were the drivers of this, made no sense for them to maintain and run two side-by-side networks.
[STEVE]: Obviously, they're sort of broadband fiber networks, which were and are the future, and they're old sort of copper-based telecoms networks.
[STEVE]: They didn't need both of them, and so were keen to obviously switch one off, and with regulator oversight, that was driven by the supply side of the industry, looking to change that redundancy, looking to remove the need for such specialist telecoms engineers for those circuit switch networks and all the aging hardware across those networks, moving to something that's new, forward-looking, multipurpose in terms of fiber-based broadband networks that demonstrate a lot greater ability to scale, greater efficiencies.
[STEVE]: This is a restructuring that's almost done.
[STEVE]: Most of the world, most of the major networks in the world, country networks in the world will have been fully decommissioned and switched off now by 2030, and many are already done.
[STEVE]: This is the oldest, the most complete, but first, and it was the driver of many of the rest in terms of the restructuring dimensions.
[STEVE]: The second leads on from it, actually, which is the restructuring from enterprise telecoms that is on-premises or equipment-based, if you like, to cloud-based services for telecoms.
[STEVE]: Now if the audience have heard of PBXs, private branch exchanges, this is the sort of switch, if you like, the telecom switch, not in the network, but that used to sit, and still to some degree does sit still in certain circumstances, in companies, offices, or sites.
[STEVE]: Every site needed one at one point, and because of that on-premises equipment, there was a cost to it.
[STEVE]: It needed to be bought, it needed to be kept and looked after and upgraded from time to time and full replacements in certain cycles.
[STEVE]: It also had constraints in that those on-site switches led to more physical, defined, and restricted service, numbers related to the office you were in, and more particularly to the desk that you sat at for a long time, and until very recently, those phones would have only rung on the desk of the number in question in the office of that PBX.
[STEVE]: That has changed a lot over the last decade.
[STEVE]: It was also accelerated by the pandemic to a good degree.
[STEVE]: Now in the enterprise world out there, over 75% of enterprises have moved away from that sort of on-premises implementation of telecoms and into the cloud with cloud-based service.
[STEVE]: It has a number of benefits for the enterprise.
[STEVE]: They're using telecoms essentially, I won't say just another data service because it is a business-critical data service, but it's another data service over the internet like so many others in the enterprise.
[STEVE]: All that equipment can be totally stripped out of all their sites around the world, leading to huge cost savings, both in terms of equipment costs and people costs who look after it.
[STEVE]: Also very importantly, particularly with the way business and the workplace has developed, accelerated by the pandemic, those phone numbers travel around with the users.
[STEVE]: They're no longer physically tied to specific desks in specific offices.
[STEVE]: The industry, as I say, is 75% of the way through this restructuring dimension.
[STEVE]: This one led to another one, and rather than the first two here, which were really led by the built big telcos in the industry, this one they joined in with a little bit more reluctantly, which is the move from standalone cloud telephony services to cloud telephony services that are integrated into broader UC, unified communications platforms, such as Microsoft Teams is probably the best known.
[STEVE]: Those big telcos, nearly all of them for many years ran more standalone cloud-based services, sometimes referred to as hosted PBX services or IP centric services, but they were standalone.
[STEVE]: It didn't combine with other things particularly, and all of that calling data, because it was a standalone external service, was outside the essential umbrella of the enterprise.
[STEVE]: That has been for the last few years, particularly again, in this case, accelerated by the COVID pandemic, has changed to services that are now fully integrated and combined with broader unified, more strategic and broader unified communications platforms, such as Teams.
[STEVE]: There are lots of advantages there.
[STEVE]: Firstly, it honors the term unified communications much better.
[STEVE]: These platforms were called unified communications, but didn't always and not often include telecoms.
[STEVE]: Telecoms was separate.
[STEVE]: When you buy Teams out of the box, you have to integrate your telecoms into it proactively.
[STEVE]: It doesn't come with telecoms out of the box.
[STEVE]: For a capability to be called unified without telecoms, it was a bit of a misnomer.
[STEVE]: It solves that missing hole, if you like.
[STEVE]: From a user's perspective, that came with a lot of benefits, therefore, because they are able to do everything under the single pane of glass of their Microsoft Teams clients and software.
[STEVE]: They can make and receive all their phone calls, as well as doing all their meetings and video from within the same interface, or the same Teams interface.
[STEVE]: And then, additionally, for the enterprise, it solves a data problem.
[STEVE]: All of that calling data then naturally flows through Teams and into that enterprise's datasets for AI and agentic leverage, so big benefits all around through this UC integration.
[STEVE]: This is a newer restructuring going on in the market, but one that's moving with real pace now.
[STEVE]: It's been going on for perhaps five or six years, this interconnections into Teams, this integration into Teams, and there are now over 100 providers that offer what's called Operator Connect, which is Microsoft's certified partnership scheme, if you like, for telcos who want to offer an integrated capability into Teams.
[STEVE]: Now, the big telcos are many of those 100 plus providers, but as I say, they got involved with this a little more reluctantly, but had to embrace it really because of the huge market pull from the enterprise world wanting to do this because the business case was so clear.
[STEVE]: But of course, there is some sort of rent sharing, some sort of margin sharing with the Microsoft piece of things.
[STEVE]: There was some value being given away to the UC provider in this particular restructuring dimension.
[STEVE]: The fourth big one, and a very dear one to LoopUp's heart, along with the other three, we embrace all the first three as well, but this one is central to our existence as a company and one that we really try and differentiate through.
[STEVE]: But it's a restructuring dimension that is now firmly in flight, particularly in the more multinational enterprise market.
[STEVE]: And this is the move from having dozens of, a patchwork of dozens of regional, often just country-specific telcos to cover your global operations.
[STEVE]: Perhaps you're using Orange in France or Telstra in Australia or a PCCW in Hong Kong.
[STEVE]: The move away from local or regional service provision, which required often dozens of telcos to cover your global operations, the move from that to a single global service provider who can offer telecommunications service domestically in all of those countries around the world.
[STEVE]: And the benefits, of course, are, I think, hopefully pretty clear, which is that level of vendor consolidation, often from tens of telcos down to one single vendor, can be such a cost saving, so beneficial in terms of time of management and operations and simplification, as well as the consistency of service provision, the consistency of implementation and the economies of scale that come with it.
[STEVE]: The fifth one, one that's absolutely going on very, very quickly, of course, like in all industries, is the influence of AI.
[STEVE]: And this doesn't just influence the telcos, it's influencing everyone in the market, moving from more reactive networks where you come across issues and you get them reported by customers, you go and investigate, you fix them to much more predictive diagnostics, the move from more statics of network optimization to a more dynamic approach and a much more personalized user experience.
[STEVE]: And wherever you look in the industry, at loop up and across all players, this is something that is essential across the board.
[STEVE]: And then to finish, one that hasn't come yet, but one that we think might come and again, is central to loop up strategy, which is the combination of two industries essentially.
[STEVE]: At the moment, you have big telcos that are into landline services and big telcos that are into mobile services.
[STEVE]: Difficult for such large entities to consolidate pragmatically into a single service provision, but the market is there for it because this distinction between landline and mobile numbers is becoming less and less because numbers ring on all devices, whatever they are.
[STEVE]: You can have a landline number, for example, a geographic number in different countries, 0203, for example, in the UK, and it doesn't just ring on your old landline phone anymore.
[STEVE]: It rings on all devices, your landline, if you still have one, your PC, your tablet, but critically, your mobile phone as well.
[STEVE]: Equally, you can get mobile numbers to ring on all devices as well.
[STEVE]: So there's a move we feel that's ready, the industry is ready for, which is to come down from two sets of vendors to one single vendor to move from two business numbers.
[STEVE]: Why do you need two business numbers to one single business number that rings and works on all your devices in the same way?
[STEVE]: And we think this is likely to be the next big restructuring dimension.
[STEVE]: That's where we see the industry having come from in recent recent years and currently, and that's why we think there's so much opportunity in this 200 year old market, the opportunity for a brand new variant of next generation telco, which covers all those bases that's VoIP based, cloud based, integrated with your UC platform, such as Teams, allows for single vendor, global consolidation across landline and mobile, leveraging AI in all its capabilities.
[STEVE]: And do please take a look at LoopUp if that proposition sounds interesting.
[STEVE]: Steve, thanks for that fantastic presentation.
[STEVE]: For everyone listening live, I'm sure you have a bunch of questions that have come up in your mind as you're hearing this presentation.
[STEVE]: In front of you, you have a Q&A box.
[STEVE]: Type in any questions that you have into that box, and we'll be sure to get to as many of those as we can on this next Q&A panel, which Steve will be joining us on in just a few minutes.
[STEVE]: Thanks, Michael.
[MICHAEL]: Steve's presentation really highlighted how the telecom industry's restructuring is opening up fresh opportunities. This live panel from our Software Oasis AI Summit features real audience questions that dive deeper into these themes. It's all about understanding how strategic pivots can unlock potential in established markets.
[MICHAEL — CHECKPOINT]: What Steve's presentation made clear is that the telecom industry's restructuring isn't just about technology—it's about strategic adaptation. It raises questions about how companies can keep pace with these changes while maintaining compliance and security. And that's exactly what this audience brought to the table in the live Q&A panel.
--- Q&A PANEL ---
[STEVE]: I hope everyone has been enjoying the event so far.
[STEVE]: I know we're just at about the halfway point, so I wanted to make sure we get to as many of the questions as we can.
[STEVE]: But to get started, I'd like to invite everybody to meet our panel.
[STEVE]: I'll start at the top.
[STEVE]: I have Frank Melke.
[STEVE]: He's the CEO over at Contraco.
[STEVE]: Welcome, Frank.
[STEVE]: Great having me.
[STEVE]: Next, we have Steve Centrifit.
[STEVE]: And Steve is the president over at SmartBridge.
[STEVE]: Welcome, Steve.
[STEVE]: Thank you.
[STEVE]: Nice to be here.
[STEVE]: Next up, Max Oglesby.
[STEVE]: He is over at Monstar Lab.
[STEVE]: Thank you, Michael.
[STEVE]: Welcome.
[STEVE]: Next, I have Akhil Varghese.
[STEVE]: He's the CEO over at Crozimo.
[STEVE]: Welcome, Akhil.
[STEVE]: Hey, everyone.
[STEVE]: Great to be here.
[STEVE]: And finally, we have Steve Flavel.
[STEVE]: He's the co-founder and CEO of LoopUp.
[STEVE]: Welcome, Steve.
[STEVE]: Thank you.
[STEVE]: Good afternoon, all.
[STEVE]: So I want to keep everybody on target, so let me just jump right into the questions that have come in up until this point.
[STEVE]: So the first question that came in was for Steve Flavel.
[STEVE]: This is from Rachel, who is located just outside of Boston, Massachusetts.
[MICHAEL]: And Rachel's question is, as you help enterprises move from on-prem PBXs to cloud-based, AI-ready telecom, what governance checkpoints do you recommend so telecom restructuring does not outpace security and compliance readiness in highly regulated regions?
[STEVE]: Yeah, I mean, regulation is an essential and extremely painful part of what we do, actually, because we're licensed in tens of countries around the world as a licensed telecoms provider.
[STEVE]: And that comes with a whole lot of regulation, administration, and effort.
[STEVE]: So it's a big part, actually, of our behind-the-scenes operations.
[STEVE]: That's one big side of it, making sure that we're doing everything according to the in-country regulator and in line with their practices.
[STEVE]: The other big side of it is the key piece that we add in to telecoms.
[STEVE]: So in addition to the multinationalness of how LoopUp deploys, which is the fundamentals of it, we do it in a way that's integrated with Microsoft Teams.
[STEVE]: Now there are different ways of doing that.
[STEVE]: There's a methodology called direct routing, which can allow, actually, for a lot of flexibility and different scenarios in terms of how you deploy your telecoms into Microsoft Teams.
[STEVE]: But there is more room for maneuver, but there's also more room for error in terms of those implementations and governance.
[STEVE]: So I think that was one of the main reasons that Microsoft introduced their second methodology called Operator Connect, which is much more locked down.
[STEVE]: It's much more governed by Microsoft.
[STEVE]: You have to pass all sorts of tests in every country that you want to offer Operator Connect in to be accredited by Microsoft.
[STEVE]: And then it comes with their accreditation and their SLAs as well.
[STEVE]: So the Operator Connect side of it is a big piece of that governance as well.
[STEVE]: Fantastic.
[STEVE]: Max, this next question came in for you, and it was from David, who's just outside of Chicago, Illinois, here in the U.S.
[MICHAEL]: And David is asking, when you frame AI as a discovery-driven, general-purpose technology rather than a SaaS product, how do you get a Fortune 500 board to shift its investment thesis from picking the right platform to owning workflows, data, and outcomes instead?
[MICHAEL]: Thanks, Michael.
[MICHAEL]: It's a great question.
[STEVE]: I think what we're advising clients to do is to take a portfolio-led approach to their AI implementation.
[STEVE]: And so at the board level, what I would imagine that might look like would be to put your AI investment in terms of almost departmental or initiative-level budgets and give those operators of those initiatives or those departments some autonomy in how they dispense with that budget or put that budget to work.
[STEVE]: I really think top-down fails in this instance, and the closer you can get to the people who are inside of the workflows in the business where AI can be leveraged to the greatest effect, the better.
[STEVE]: And so I think you sort of have to set up an environment of permission and trust and get that decision-making sort of out of the boardroom, frankly, and down into the lower levels of the company where I think the people actually have the insight and maybe the inspiration to know what to do with that opportunity.
[STEVE]: This next question is for Steve Centrifit.
[STEVE]: And Steve, this comes to us from Olivia, who is in Columbus, Ohio.
[MICHAEL]: And Olivia is asking, in your client work, what has been the most effective way to blend predictive models, generative AI, and agents into a single workflow without creating unmanageable spiderwebs of integrations?
[MICHAEL]: That's a great question.
[STEVE]: So I mean, one thing we look at is, does the organization have a common data foundation where you're not, we don't ever recommend you tie your AI directly into your operational systems.
[STEVE]: Put a data foundation in that allows the data to sit kind of in a data lake or something like that, where the AI can then ingest and make the actions.
[STEVE]: The other aspect of this, and to Olivia's question, is we recommend, especially getting started early in your AI journey, is to look at more human in the loop from a model standpoint, where your generative AI is making recommendations, but it's still up to the individual, the predictive model.
[STEVE]: So you could have machine learning models running in the background, but all of it's being served up in a combination of a generative AI solution that the human is having interactions with.
[STEVE]: They're still making the decisions, making the actions.
[STEVE]: So those same intelligence from a predictive model can then also be fed into analytical tools like Power BI, Tableau, and others that allow the entire organization to see, have that visibility.
[STEVE]: The next question comes to us, this is for Frank, from Jonathan here in the US in San Francisco.
[MICHAEL]: And Jonathan is asking, when you warn about AI as a multiplier of broken processes, what is the first step or diagnostic or approach you use with executive teams to surface race to the bottom patterns before they hardwire them in to a large scale AI role?
[MICHAEL]: The first question we always ask, do we have the right data for the right problem?
[STEVE]: And this is quite important that even the board understands if we have bad data, we are only worsening these data.
[STEVE]: If we have a bad process in place, the process will be much quicker, but not in a higher quality.
[MICHAEL]: So the very first question is, what's the problem?
[MICHAEL]: Have we all a common understanding about the problem?
[MICHAEL]: And do we have the right data to solve the problem?
[MICHAEL]: Sounds like an effective approach.
[MICHAEL]: The next question is for Akhil, and this comes to us from Anthony, who is right in New York City.
[MICHAEL]: Anthony is asking, when you talk about deeply connecting AI to CRMs, calendars, order systems, and entitlements, what architectural patterns do you recommend so these agents stay maintainable as vendors, schemas, and policies continue to change over time?
[MICHAEL]: Hey, Anthony.
[MICHAEL]: Yeah, great question.
[STEVE]: So I think we kind of, there's a fairly extensive AI governance setup that goes in before any of these integrations happen, and they mainly revolve around defining every single element of what data exists, what tools exist, what calls exist, and implementing the guardrails and the logging requirements, the auditing requirements, the human oversight requirements on all of those data systems.
[STEVE]: Any modification of those systems over time, whether that's, you know, obviously not simply the addition of new data in the same format, but any changes in the format, any changes in the structure, you need to set really rigorous rules as to how you also have to simultaneously update the rules of engagement for AI with all of these systems and practices.
[STEVE]: So taking the time to not only set up the structure, the guardrails, the rules first, but then also define how these rules will evolve as things change, and then enforcing that in a completely deterministic manner, just essentially not allowing people to make changes until they've also updated the associated, like, rules of engagement for AI can kind of protect you as you add agents, and then your team can experiment quite liberally, which is how you should experiment with AI.
[STEVE]: You should be quite liberal with the experimentation, but strict with the governance.
[STEVE]: And the next question comes to us for Steve Flevel, and this is from Andre, who is in Minneapolis, Minnesota.
[MICHAEL]: And Andre is asking, when global carriers retire PSTN and copper, what are the most common pitfalls you see in multinational rollouts, especially around numbering plans and local emergency calling, and how do you de-risk those early in the process?
[STEVE]: Well, numbering is one of the biggest headaches of telecoms when you're in that sort of change of phase, you know, the whole porting process over to new suppliers is fraught with difficulties because there's always a losing carrier in the situation, so, you know, there's somebody you want to move service over to, and there's somebody you're taking service away from, and there's never a particularly good incentive for the losing carrier to easily accommodate that, shall we say.
[STEVE]: So that's one of the common pitfalls is porting, you know, something that we've worked hard on in that respect is to really codify the requirements of every carrier that we port away from and over to so that we know exactly the structure, not only the nature of the data, but the structure of the data that we have to get from a customer in order for that shift to happen.
[STEVE]: So it's one of the main drivers of a seamless switchover and implementation, particularly when you're moving over to a provider like LoopUp that can put everything through teams.
[STEVE]: I think there's so much to be gained that, you know, you've just got to manage the process well.
[STEVE]: We have a team of consultants that will help the customer through that process, a whole team associated with number porting that will work with the carriers involved.
[STEVE]: So I think it's all in the process and the onboarding.
[STEVE]: Makes sense.
[STEVE]: And these are not small undertakings, so having the right support obviously is key.
[STEVE]: Max, the next question comes for you from Nicole, who is in Alpharetta, Georgia, I believe right outside of Atlanta.
[STEVE]: And Nicole is asking, you cite research showing most gen AI pilots fail to deliver P&L impact.
[MICHAEL]: What is the first financial signal you ask CFOs to track that proves an AI initiative is becoming an operating model shift, not just a lab experiment?
[STEVE]: I mean, I think the classic two would be some increase in revenues or some decrease in costs.
[STEVE]: That would be sort of the classic view.
[STEVE]: I think there's a lot of instrumentation that you could do sort of upstream of that, though, around the human inputs to various processes and whether those either human inputs are becoming more efficient or the quality of the outputs are becoming measurably better.
[STEVE]: So it really depends what workflow it is.
[STEVE]: To give you just a quick case study, we're doing work with a civil engineering firm, and it's really clear.
[STEVE]: There's a certain process in their kind of step-by-step land development sequence that took three to four weeks in the old way of doing things, quote-unquote.
[STEVE]: And with some AI-enabled automations that we've built for them, they can do that same amount of work in half a day, roughly, so from three or four weeks to half a day.
[STEVE]: That's a pretty clear signal.
[STEVE]: Not all of these are going to be that easy to measure, but I really think it comes down to understanding what that specific workflow is.
[STEVE]: And I think you'd find that a lot of the KPIs are similar to what you would have found in any kind of classical orchestration or automation job.
[STEVE]: It's not really new ground that you're trying to break from a measurement standpoint.
[STEVE]: I think it's just the tools are evolved.
[STEVE]: This next question is for Steve Centervit.
[MICHAEL]: This comes to us from Monica, who's in Houston, Texas, and Monica asks, for enterprises that want to move from rule-based automation to agentic AI, what are the concrete criteria you use to decide which steps in a process can be safely handled or handed to an AI agent versus kept strictly under human control?
[STEVE]: Yeah, it's a, Monica, I'm in Houston as well, so it's nice to get a question from the hometown.
[STEVE]: So and it is a great question.
[STEVE]: One of the things that we always look at when we're starting that process is what's the risk level.
[STEVE]: So we work a lot in the oil and gas industry.
[STEVE]: Safety is a big concern.
[STEVE]: So when you're looking at automation, whether you're looking at traditional RPA, just robotic process automation, or you're moving more into the agentic AI, really the risk profile, you have to look at the risk profile of your workflow and understanding what's the safety net if an agentic AI agent makes a decision and does something.
[MICHAEL]: How quickly can you revert back?
[MICHAEL]: And that really comes down to a risk management approach that you have to apply to that workflow.
[STEVE]: We recommend organizations start on a conservative nature when you start going down that path and save the automation of those critical steps and keep those to the lower risk items, but it all comes down to risk management and understanding how to measure that risk.
[STEVE]: This next question is one that came in a few times from a few different people, so I'll pick one here.
[STEVE]: This is for Frank, and Frank, this is from Jonas, who's in Berlin, Germany.
[MICHAEL]: And Jonas is asking, when a global enterprise is still running on mainframes across multiple jurisdictions, how do you sequence modernization and AI adoption so they avoid the illusion that AI alone can make legacy architectures go away?
[STEVE]: That's a very interesting and good question, and it's one we had to answer with one of our clients recently, a big multinational company with mainframe databases, very modern technology facing to the web, but in the back there was the mainframe holding more or less data for 60 years, all client transactions.
[STEVE]: And the idea was AI is the silver bullet, and you can offload all the data from that mainframe into a modern data lake architecture.
[STEVE]: Yeah, you have very structured data in that mainframe, but on the other hand side, in all the other systems, you had 47 client IDs for one and the same person, and there is no AI approach out there who can automatically somehow bring these data together and save it into a data lake.
[STEVE]: That brings me back to my first question.
[STEVE]: First, you need to understand the problem, then you need to understand what data you have in place.
[STEVE]: That is human beings needing to answer that question, and then you can ask AI to help you to merge this data.
[STEVE]: Yeah, these are real world problems that many executives are dealing with, and there is no black and white, one size fits all answer.
[STEVE]: This next question is for Akil, and this comes to us from Lauren, who's in, let's see, she's in Phoenix, Arizona, and Lauren is asking, you highlight the risk of celebrating higher connection rates that actually degrade downstream close rates.
[MICHAEL]: How do you instrument funnels so leaders can see in real time whether AI is improving or diluting overall pipeline quality?
[MICHAEL]: Yeah, great question, Lauren.
[STEVE]: I think that what I have found most effective in doing this is to, and I think Max alluded to this as well in what he said, is define the clear metric that moves the needle and make sure that that metric is measurable.
[STEVE]: So for example, if you're trying to automate outbound sales, you could say that the metric is response rate, but I have personally found that response rate is kind of a terrible metric because you could cold email a lot of people and say, hey, I'm offering you a free trial of something and there's no risk and all that'll happen is we give you this much service for free.
[MICHAEL]: And you may get a high response rate, but that's not ultimately relevant to the net result, which is do you close someone that is actually going to pay your business, right?
[MICHAEL]: So what I have found really useful is to look at anything that you're trying to automate from the perspective of what is it net for you currently in terms of how does it move the needle on your revenue?
[MICHAEL]: How much does it cost you in terms of the people that are doing it currently or the software that you're using to do it currently?
[MICHAEL]: And then you will be better able to identify exactly what metric you should measure when you're evaluating how successful an automation is, which is usually the measure you should use when you're deciding how to roll out an automation.
[MICHAEL]: So if you're converting 6% of the people that you reach out to, you have to realize that if that drops to 5% or 4% and your revenue is really high, even if your AI automated solution is almost free and your salespeople cost $50,000, $100,000 a year, you may still be losing money because that 2% drop may actually net less revenue.
[MICHAEL]: So it's really important to define the criteria for success you have beforehand and ensure that criteria is measurable.
[STEVE]: And ideally, try to set all of that in stone before you begin the AI deployment because I found that once people start deploying, they twist and turn just to make it seem like the AI is doing well.
[STEVE]: This next question comes to us for Steve Plevel.
[STEVE]: This is from Raleigh, North Carolina, from Grace.
[MICHAEL]: Grace's question is, for an organization like ours, standardizing on Microsoft Teams calling over 100 countries, how do you advise them to balance centralization with the need for regional resiliency and local carrier redundancy?
[MICHAEL]: You have to have both, so you don't have to have both, but it would be ideal to have both.
[STEVE]: And the way our business model is all about the consolidation side, of course, and all the efficiencies you get from that.
[STEVE]: So you only have one contract, one support team to deal with, it will be there all around the world, one renewal process, one portal for all your analytics and management of all your people.
[STEVE]: There's so many efficiencies to the consolidation.
[STEVE]: The way we deal with the challenge, which is the regional optimization, is underlying loop up, we have 25 carrier partners and growing all the time that we work with, and we choose them on the basis of best in class for their region and their local geographic situation.
[STEVE]: So we're getting that local and regional optimization through the underlying carrier partners that we work with.
[STEVE]: So you can have the best of both worlds in that respect.
[STEVE]: Also we have a sort of a backbone of network ourselves, so rather than the normal situation, which is if you have an international call, that may get handed off between carriers multiple times before it gets to its end destination.
[STEVE]: Because of the core loop up backbone and the various data centers we have around the world that are all interconnected, we only have one handoff to the terminating carrier locally.
[STEVE]: So you're getting better quality, you're getting lower cost, and you can have your cake and eat it in that respect.
[STEVE]: This next question comes to us for Max, and this is from Helena, who's in Stockholm, Sweden.
[STEVE]: Helena's asking, you compare AI to prior general purpose technologies like electricity and the internet combustion engine, the internal combustion engine.
[MICHAEL]: What is the modern equivalent of rewiring the factory that digital leaders should be funded over the next 24 months?
[MICHAEL]: Thanks, Helena.
[STEVE]: It's funny, there's a book by a guy named Wolfgang Schivelbach, I think is his name.
[STEVE]: He has like the funniest name, great name.
[STEVE]: And he talks about the earliest stages of the industrial revolution and power in factories.
[STEVE]: And the very first thing that people did with the rotary engine, the steam engine, was to pump water back up mill chases.
[STEVE]: So the factories stayed hydropowered.
[STEVE]: But if there was a drought or the water stopped flowing, they would use a motor to pump water up to the top of the mill chase and kind of force it back through the mill wheel.
[STEVE]: I think we're at that stage of AI right now, where we're effectively using AI as a sort of engine along the side of our typical enterprise workflows.
[STEVE]: We're using AI to metaphorically pump water up the mill chase.
[STEVE]: I think you're only seeing the beginnings of AI operations truly embedded in enterprises.
[STEVE]: So I don't know if I'm answering Helena's question quite correctly, but if the request was to compare the current sort of AI moment to previous revolutions, I'd say we're in the first half of the first chapter, where we're basically using AI as a way to augment legacy systems and legacy workflows.
[STEVE]: But I think all of that will change over the next 5, 10, 15, 20 years.
[STEVE]: And future enterprises, I think, will be sort of unrecognizable compared to what we're doing now.
[STEVE]: Yeah.
[STEVE]: Even though we're deep in the middle of where we are, it's still early on.
[STEVE]: This next question is for Steve Senterfit.
[STEVE]: And it comes to us from Brian, who's in Charlotte, North Carolina.
[MICHAEL]: And Brian is asking, when you introduce agentic AI into existing BPA estates, how do you prevent shadow automations from popping up in the business and instead keep everything traceable with clear audit trails and ownership?
[STEVE]: I appreciate the question.
[STEVE]: There's two aspects of this.
[STEVE]: One is around the governance that the organizations need to put in place.
[STEVE]: That's important.
[STEVE]: That will also help reduce shadow.
[STEVE]: Just like you had shadow IT, you wind up having shadow automations popping up and things being done outside of it.
[STEVE]: So that's a piece of it, making sure that you have the right governance framework in place from an operational standpoint.
[STEVE]: The other, and he referenced it in the question itself, and that's the audit and traceability.
[STEVE]: Knowing that you have that and that it's embedded in the feedback loop and the transparency in the solution.
[STEVE]: Whether you're putting analytics on top of that, so the organization can see the decisions that have been made, the actions that have been taken by the AI solution.
[STEVE]: That also is needed for ongoing training of those agentic AI agents, especially if they're being driven off of any type of machine learning algorithms and data sets that are being generated by that.
[STEVE]: Those would be the two main pieces.
[STEVE]: The other piece is on the human side.
[STEVE]: And that's ensuring that when you're implementing these, you're not layering the AI agentic agents just on top of a workflow, but embedded in the workflow.
[STEVE]: The reason that's important is if it's just layered on top, one, the adoption will typically lag.
[STEVE]: And I think that's one of the reasons why a lot of enterprises don't see pilots scaling up is they're layering on top of an existing workflow versus embedding it in maybe a re-engineered workflow that will also help prevent shadow processes popping up.
[STEVE]: This next question is for Frank, and this comes to us from Eric, who's just outside of Boston, Massachusetts.
[STEVE]: And Eric is asking, you emphasize diagnosing before prescribing.
[MICHAEL]: What does a concrete initial discovery sprint look like when you were deciding whether AI is the right lever versus simpler processes or policy changes?
[MICHAEL]: Very complex question, as the environment is complex.
[MICHAEL]: The first question is, what is the department or the departments involved in that problem?
[MICHAEL]: You need the human side, and you need to understand how that process looks like and what problem needs to be solved.
[MICHAEL]: Rather, the discovery will be more on the use case than on the technology.
[MICHAEL]: And if you understood the use case, then you can recommend the best approach from a technology point of view.
[MICHAEL]: For sure, you need to consider some frameworks.
[MICHAEL]: You need to consider legal aspects to it.
[MICHAEL]: But first, you need to understand the use case, and then you can recommend.
[STEVE]: And I want to keep everybody on track for the next session.
[STEVE]: But I'm going to pose the final question to Akil.
[STEVE]: And this comes to us from Kevin, who's in Seattle, Washington.
[STEVE]: And I'm picking this question because I think it's something that applies to just about every organization at the event today.
[MICHAEL]: So Akil, Kevin is asking, when you map an end-to-end sales or support workflow, what is the most overlooked system or tool integration that tends to make or break whether an AI agent can operate as a true teammate rather than a disconnected bot?
[STEVE]: I have a definite answer to this because it's something that I think everyone overlooks.
[STEVE]: So generally with support sales cases, people think, okay, you're going to need access to the orders, you're going to need access to the policies, you're going to need access to the FAQs.
[STEVE]: And all of that is obviously essential, but no one really misses that.
[STEVE]: One thing that we found when we implemented these systems is that in many companies, especially the smaller ones, sales and customer support also have a direct line to engineering.
[STEVE]: So they very often talk to engineering directly and say, hey, this many customers want this feature.
[STEVE]: I've seen this pattern many times.
[MICHAEL]: Is this something we can do?
[MICHAEL]: And if so, when can we promise this?
[MICHAEL]: And can this escalate?
[MICHAEL]: And often when people automate AI systems to do sales and customer service, they very rarely have that integration.
[MICHAEL]: So it can either be a direct integration with JIRA or whatever ticket system you use, or just a way to contact the engineers and interpret their responses on dates and estimates and things that they can commit to as autonomous agents.
[MICHAEL]: So yeah, I think integrating with leadership, management, engineering, the decision makers in determining like, hey, even if we can't do what someone's asking right now, is there a way that we could?
[MICHAEL]: And especially with AI, when it handles all of your queries and has a better sense of what people are asking for at scale, it's really, really good at prioritizing what engineers should prioritize within the business.
[MICHAEL]: Fantastic.
[STEVE]: Well, I think this has been a great discussion.
[STEVE]: Frank, Steve, Max, Akil, Steve, I'd like to thank everybody for joining us at the event today and the fantastic presentations and conversation.
[STEVE]: I know that for everyone that's listening in live, you may have some questions that we didn't get to.
[STEVE]: Up at the top of the site, you'll see the big orange button.
[STEVE]: If you click on that, we'll get you in touch with any of the speakers that you've heard on this panel or throughout the event, as well as, you know, you can also reach out to anybody on the team here.
[STEVE]: But with that, I am going to get everybody on their way to the next presentation, which is from Daniel Korolowski over at TLV Tech.
[STEVE]: So thank you, everyone, for joining us on this panel.
[STEVE]: Thank you.
[STEVE]: Appreciate it.
[MICHAEL]: Today, we learned how Steve Flavell and LoopUp navigated massive industry shifts to transform their business model. Reinventing from conference calls to a global telco wasn't easy, but it shows the power of strategic pivots. So, thank you, Steve, for sharing your journey with us. If you'd like to connect with Steve, you can find him on LinkedIn. And if you haven't already, come join us over at softwareoasis.com — that's where all of this really comes together. Don't forget to subscribe and maybe even leave us a review if you enjoyed today's episode. Thanks for listening, and I'll catch you next time.