DR: [00:00] That’s what’s really interesting about the ontology. You don’t necessarily have to model the whole telco. You just do this little piece and then you add onto it. It just keeps growing and growing as you come up with new ideas, have a new question, want new information.
David Haselwood: [00:14] But it still works with their old systems. That’s the important part.
DR: [00:16] Totally.
David Haselwood: [00:17] That’s the key thing is you’re just adding on top of it. Okay, that sounds pretty awesome.
DR: [00:21] It’s super cool.
Announcer: [00:28] This is “Telco in 20,” a podcast that helps telco execs achieve a competitive advantage with AI and the public cloud. It is hosted by Danielle Rios, also known as DR.
[00:38] For this episode, David Haselwood, a longtime friend of “Telco in 20,” is interviewing DR as part of our “What’s up with Totogi” series.
DR: [00:47] Hey, guys, it’s time for another “What’s up with Totogi.”
Music: [00:51] What’s up with that? What’s up with that? What’s up with Totogi?
DR: [01:02] What’s up? MWC is four weeks away, and I can’t wait to share Totogi’s BSS Magic with everyone at the Fira. Operators are already deploying it live in production at Tier-1 and Tier-2 telcos across the globe. But with any new product that solves problems in a new way, we get lots of questions. Like, what’s an ontology? Is BSS Magic actually a BSS? And what was up with that whole Düsseldorf thing with 600,000 lines of AI generated code? I get why the industry is so curious. The telco world is flooded with AI vendors slinging buzzwords, all hype, zero delivery. So, when something actually works and delivers real revenue impact, everyone wants to know more. So, today on the podcast, we’re flipping the script.
[01:49] I’ve invited my friend David Haselwood to interview me about BSS Magic. We’re going to dive into what a breakthrough biology project from the 1990s can teach telcos about making AI actually work. Why semantic inconsistency is the real reason telcos are struggling to implement AI and how Totogi is working with Tier-1 operators to put AI into production. So, let’s take 20.
[02:17] Hi, David. Welcome back to “Telco in 20.” I’m so excited to have you on the podcast.
David Haselwood: [02:22] Hey, DR. I’m excited to be here.
DR: [02:23] Well, you’re my good friend. We’re going to do something a little bit different today for this special “What’s up with Totogi.” You’re going to be asking me the questions.
David Haselwood: [02:31] I love it. I’m excited. There’s one thing I want to talk about. I thought about you the other day because this person we both follow said something about how bragging about AI in 2026 is like bragging that you have cloud storage. Just a complete garbage throwaway statement. And since you are the queen of the cloud, I thought of you. And so, I don’t know what your thoughts about that, where people talk about AI, but there’s no value, nothing going on with it.
DR: [03:01] It’s so funny because I saw that too. It came from Lulu Cheng Meservey, this brilliant PR lady. But there’s the second part of the tweet where she said, “There’s no more value coming from just listing a bunch of AI buzzwords.” People are super tired and over it. And so, really what vendors need to do is they need to skip to, what is the thing that it actually does and then prove to me that it actually does that.
David Haselwood: [03:22] Right, exactly.
DR: [03:23] Yeah.
David Haselwood: [03:23] So, I know something that you’re going to be talking about a lot that I’m coming up to speed on is BSS Magic.
DR: [03:29] Yeah.
David Haselwood: [03:30] I’m doing the best I can to come up to speed on it. I’m a little bit confused. Walk me through what is BSS Magic? What do people get right about it? What do people not understand about it?
DR: [03:39] Well, I think it is confusing and so I appreciate that. You’re always so great with me in giving me feedback. But I remember that you have a degree in natural sciences from Johns Hopkins.
David Haselwood: [03:49] I do.
DR: [03:49] And so I’m going to use a biology example.
David Haselwood: [03:51] Great.
DR: [03:52] So, there’s this thing called the gene ontology.
David Haselwood: [03:54] Yeah.
DR: [03:55] In the 1990s, biologists had thousands of databases describing genes and proteins. None of them spoke the same language. Everyone was calling the same thing, something different. One database described a DNA sequence one way, another one used a completely different terminology, no shared language. And so they couldn’t reason across the data sets because identical concepts had just different definitions. And so, they built something they called a shared semantic layer, 44,000 concepts, standardized definitions. Now the databases connect through one common language. The breakthrough wasn’t about building a better database. It was about building what I’ve been talking about a lot, semantic consistency. And so, telcos have this problem big time inside their businesses. They’re running hundreds of systems where subscriber means something different in billing, than CRM, than care, than provisioning. And so, yes, APIs can move the data, but they don’t really resolve meaning and create comprehension.
David Haselwood: [04:53] Yeah.
DR: [04:53] And so, BSS Magic is the gene ontology for telco. There’s three layers to our product. There’s this data layer that connects existing already installed systems. The second layer, which is the ontology, that semantic consistency, establishing shared definitions and models, the telco’s business processes. And then the top layer, the AI layer helps you do whatever you can do with AI. Build applications, deploy agents, gets all its context from the ontology. And so, biology took 20 years to get there. We’re handing it to telcos prebuilt on day one. And I think this is the big unlock for AI in telco.
David Haselwood: [05:30] That makes sense. I understand what you’re talking about with the genes and you had tons of people sequencing genes and they couldn’t talk about what the difference between them were. And so, you had all this interesting stuff going on and you could not enable anything. So what you’re saying is with this approach, you’re able to enable AI with it, correct?
DR: [05:48] Well, yeah. It’s giving everything in telco the same context and the same meaning. And once everything has been normalized, then AI now has the context to not hallucinate, have those guardrails, understand what you mean, and really enable regular people, executives inside the telco to converse with their systems, to converse with their data and not have a consultant go off and code it up and say, “Well, this equals that in these two different systems and here’s the report.”
David Haselwood: [06:18] Preventing the garbage in, garbage out of AI.
DR: [06:20] Totally.
David Haselwood: [06:21] Okay, got it. So, the BSS Magic, it’s actually not a BSS.
DR: [06:27] Yeah. I think a lot of people think when they hear BSS, they think, “Oh, it’s a bunch of modules. It’s a suite of products.” It’s not.
David Haselwood: [06:35] Okay.
DR: [06:36] And so, I think that’s probably the biggest misconception about the product.
David Haselwood: [06:39] That is important. Okay, got it. So, the project you guys did at that conference in Germany where you put some guy on a desk and he built a bunch of stuff.
DR: [06:49] Totally.
David Haselwood: [06:50] That was BSS Magic, right?
DR: [06:51] Yeah.
David Haselwood: [06:51] Walk me through that, because I watched that. I didn’t quite understand what was going on there.
DR: [06:54] You didn’t know about all my press releases and my bragging. But yeah, you’re talking about Düsseldorf and the AI-Native Telco conference that I was at. And we really wanted to highlight that AI coding capability is insane and off the charts and you can vibe code a telco app provided you give it a bunch of context.
David Haselwood: [07:15] Right.
DR: [07:15] And that’s what our ontology can do. So, BSS Magic has an ontology based on TM Forum standards. This is an industry body. And when you give it like, “Hey, go build this module.” It’s like, “Oh, I know exactly what you’re talking about because I know everything about telco.” So, I’m going to use the power of Claude Code and Opus 4.5 is just off the charts. It now reaches human capability of software. Combine that with an amazing ontology and what you want to build, and you can build fricking anything. And we’ve generated 600,000 lines of code in eight hours. I think today we could do it in about two hours and we generated six clean new modules. I don’t think that’s what you would use it for, but it was just a demonstration of the power of AI with an awesome ontology and telling it what to do and it’s just super powerful.
David Haselwood: [08:04] And then you had them add a couple of extra audience features on there. Was that to just show that this was not a canned thing?
DR: [08:10] Totally. I do think telco executives are just sick of AI slop, vendors coming in saying we can do it, paying them, it doesn’t work, the promise isn’t there. And we’re like, “How do we prove that this really works?” And so, we were showing the code being generated. We brought a forward deployed engineer, we asked for real features. We had this one feature where the telco exec walked up and said, “Can you send me an SMS that I’ve been onboarded?” We had to think about how we were going to do it. 20 minutes later, we sent the guy an SMS. We’re like, “What do we need to do to prove that this is fricking real?” That’s why we did it. So, you’re exactly right.
David Haselwood: [08:48] I think now I understand what BSS Magic is, but let’s turn it around and think more about the customer.
DR: [08:55] Yeah, sure.
David Haselwood: [08:55] Because at the end of the day, you have to be solving some need of theirs.
DR: [08:58] Totally.
David Haselwood: [08:59] What’s the biggest pain point that you’re trying to fix? And maybe it’s in the form of different things people have come up with because I get the whole structure, but at the end of the day, someone’s going to want to be like, “Can you do this?”
DR: [09:11] Yeah. I think everyone’s like, “What’s the use case you solve?” And very simply, it’s the semantic inconsistency problem. If you looked inside these telcos, you’d be shocked at how awful they are. The state of legacy spaghetti code they have going on is insane. It’s like 1990s over there. Hundreds of applications all integrated to each other, all with different data models. Keeping all of this straight has been embedded in integrations and business logic and customization, true fact. They spend 80% of their IT budgets maintaining this and about 20% on innovation. So, when you call on your telco and they can’t answer basic questions about your account, it’s because the poor person on the other side of the line is trying to figure out what system to access to find the information that you’re asking. It’s just a disaster.
[09:58] And so, what if you had a tool that solved this for you once and for all? You didn’t have to always pay a consultant to do that translation for you to get you the information you were looking for in your company. And so, if you had a tool that abstracted that all away and you could just talk to your systems, I’m not talking about a business warehouse or a data lake that’s historical looking back data. I think it’s more than a knowledge graph, but we’re getting closer, giving context and meaning to the entities and the objects inside of your telco, but an executable callable ontology that’s a digital representation of your business.
[10:33] And as you attach more systems and more information, it compounds in value. You could ask it anything you want. It’d be hugely, hugely powerful. Telcos need insights in their business. I think the big pain point is semantic inconsistency and the solution is BSS Magic.
David Haselwood: [10:49] So, you’re talking about the insights looking into finding real problems that they’ve got. So, let’s get to the point, how do you know this works?
DR: [10:55] Yeah. I think the best way you know that it really works is telcos are really using it and deploying it to production.
David Haselwood: [11:00] Right.
DR: [11:01] I think there’s a lot of telcos out there that are hiring big names, IBM, Accenture, Deloitte, all these consultancies that are coming in and selling big AI projects and where’s the beef? They’re not really delivering any value.
David Haselwood: [11:15] Okay.
DR: [11:16] And so, for Totogi and BSS Magic, I think we’re close to 10 Tier-1s that are starting to build out their ontologies with BSS Magic. And so, I’ll tell you a story. One of my favorite ones, this is a large group in the Middle East and Africa. They had a problem, which is something they call dormant cells. These cell towers just quietly fail. They’re not alarming on a dashboard. Nothing seems to be wrong, but customers are calling in with complaints of loss of service. And so, they obviously are losing revenue when traffic isn’t flowing through their network. And certainly risks churn because customers were like, “Your network sucks.” So, they wanted to see if we could help them detect this. It required connecting to all these different systems, Huawei, Ericsson, their BSS, their field service scheduling, customer communications.
[12:02] These systems are integrated. They’re passing data between each other, but they’ve never connected the information in this way. So, to solve this problem previously, it would take a field tech person about 48 hours to go figure out what was the problem. We took that time down to 30 minutes and now we’ve added a machine learning script to it that can predict when it’s going to happen. So, now we can stop it from happening in the first place. So, real-world situation, real revenue impact, deployed to production. I can’t think of anything that shows this really works then actually proving it’s providing value.
David Haselwood: [12:38] And they could tell how much revenue they were losing?
DR: [12:40] Totally.
David Haselwood: [12:41] Get out of here.
DR: [12:41] Yeah.
David Haselwood: [12:42] There’s the churn part of it. There’s the truck roll when they got to go out there and do it. There’s all those different components. Okay. So, you guys were able to figure all that out and they could see how much they were going to be saving up then.
DR: [12:52] Yeah. You can tie revenue to a cell. And so, let’s say you had 25 dormant cells, but you only have the staff to fix 10. They were like, “Give us the prioritized list of which ones we should fix in order based on revenue.” And so, they’d never really looked at it in this way. Sure, you can guess, a highly populated area probably has a lot more revenue through it, but give us a list and we’re like, “Okay, here’s your prioritized list for field service to go out and fix the dormant cell.”
David Haselwood: [13:17] So, that’s where you took their data, their systems, you added the BSS Magic layer. And what you were saying there at the top, that sounds like the AI part where you were actually doing the predicting and stuff like that.
DR: [13:27] Yep. So, we built the ontology, how the objects all related. Like you said, bottom layer connected the systems, modeled their ontology, how they all relate, how do you detect a dormant cell, and then used AI vibe coding to generate a dashboard. And they’re like, “Great, we want to put this in our customer care group so they have a live dashboard of dormant cells so when someone calls in, they know, oh, you’re on the dormant cell list. Done.”
David Haselwood: [13:51] Yeah.
DR: [13:51] They’re like, “Can you send an SMS to tell them we know?” And we’re like, “Yep.” So, we added that. And so that’s what’s really interesting about the ontology is that you don’t necessarily have to model the whole telco. You just do this little piece and then you add on to it. It just keeps growing and growing as you come up with new ideas, have a new question, want new information.
David Haselwood: [14:10] But it still works with their old systems though. That’s the important part.
DR: [14:12] Totally.
David Haselwood: [14:13] That’s the key thing is you’re just adding on top of it. Okay, that sounds pretty awesome.
DR: [14:15] It’s super cool. So, I think a lot of people think, “Oh, let’s consolidate on a bigger vendor so I have more modules from the same person.” And I’m like, “The root cause is the semantic inconsistency.” And you think that vendor consolidation helps you because it’s one data model. But in a telco, there’s no one vendor that does the whole thing end to end.
David Haselwood: [14:35] Yeah. As long as more than one vendor, you’re going to have the same problem.
DR: [14:38] Correct.
David Haselwood: [14:39] So you have to solve that problem for now and for always. Okay.
DR: [14:41] Exactly.
David Haselwood: [14:42] I get it. The cell tower one, that’s a pretty good story. That’s going to be pretty tough to beat because I really like to figure out how you’re solving these problems that actually helps me understand how the product works, how it’s rolled out, how the customer uses it.
DR: [14:55] Yeah. I could talk about this all day long. We had another customer in Asia. They were one of the first customers trying out BSS Magic and they gave us a super easy one. They’re like, “Hey, our enterprise sales reps take five minutes to do this process. Can you take it down to two minutes?” They’re like, “We tried this with some horizontal AI vendors and they can’t get it to work, but maybe this ontology thing, this telco context.” And so, totally, we did it. It was super easy and we cut the time down to one minute.
David Haselwood: [15:20] Yeah.
DR: [15:21] But when they saw what we did in Düsseldorf, they were like, “Hey, that’s really, really interesting. And we’ve been thinking about starting to reduce our vendor dependence, build our own BSS modules, but we were going to code them ourselves. Can we use BSS Magic to create modules?” And we’re like, “You can.” And so, again, it’s the capability of AI now with an ontology and context, incredibly powerful. And so, I’m super excited to talk about this and share these stories at MWC.
David Haselwood: [15:50] You are the de facto queen of MWC.
DR: [15:53] I am.
David Haselwood: [15:54] There’s anybody that I wanted to talk to about what was going to happen there. So, as you look ahead, what are telcos exactly looking for at MWC this year?
DR: [16:02] Well, I think it takes us back to the conversation that you started at the top with that tweet from Lulu about people are tired of the hype and the buzzwords and they really want to know if something works and it can deliver real value. And I think Totogi has that and we have the stories and we have the customers. And so, I think everyone should cruise by Hall 2. It’s going to be super epic and come visit Totogi.
David Haselwood: [16:25] That sounds fantastic. I’m really excited about that.
DR: [16:28] Well, David, I appreciate you coming on the podcast and asking me all these awesome questions about BSS Magic.
David Haselwood: [16:34] I love learning what you’re doing. Thanks for having me on. It’s so exciting.
DR: [16:37] Awesome. Yay.
[16:42] Stick around. We end each podcast with a “Telco in 20” takeaway. I’ve got two minutes to tell you something you need to know.
[16:53] David asked me a great question. What’s the biggest pain point BSS Magic solves? Here it is. Your systems don’t speak the same language. Inside your telco, subscriber means something different in billing than it does in CRM, care or provisioning. These systems pass data between each other, but they don’t actually understand each other. So, what do you do? You pay consultants to make the translation. Every integration, every API, every report, someone has to make sure that information makes sense as it moves from system to system. AI without context hallucinates. It gives you answers that sound right, but aren’t. Your systems can’t give AI what it needs to actually understand your business. Here’s what the industry gets wrong. They think connecting systems via APIs solves this. It doesn’t. APIs move data. They don’t translate meaning.
[17:49] You’re just passing information faster. It’s still gibberish on the other end. That’s why you can have everything integrated and still not know how many active subscribers you have. You don’t have an integration problem. You have a language problem. The answer isn’t connecting your systems to each other. It’s connecting them all to something that actually understands what they mean. BSS Magic is the missing layer. It’s a living model of how your business really works. And when AI has that, it stops guessing and starts knowing.
[18:21] We’re deploying this in production with Tier-1 operators right now. Real use cases, real revenue. Solve the language problem first, then AI actually works. I’m breaking all of this down in one of my must-see talks dropping February 17th on YouTube. That’s just two weeks away, so be on the lookout. Until then, tune to more “Telco in 20” episodes. Like and follow and leave us a five-star review. Be sure to connect with me on LinkedIn and X @TelcoDR. Sign up for my game-changing email newsletter on telcodr.com and check out our amazing YouTube channel and hit that subscribe button. Later nerds.
