Speaking of TM Forum’s DTW, I can’t wait to share the amazing results of Totogi’s Catalyst project championed by BT and with help from delivery partners CloudBlue and Infosys. The project is called “Accelerating Dynamic Network Marketplaces,” and it demonstrates how telcos can use a Totogi BSS Magic AI agent to radically simplify the process of placing a complicated order, like a configured multi-site VPN. Today, it requires a maze of APIs, manual processes, and inter-team handoffs. Our project uses AI to streamline the journey, cutting quote-to-order time from days to minutes. Be sure to check it out!
In other exciting news, our Telco in 20 podcast won GOLD! We're thrilled to have received a 2025 Gold Hermes Creative Award for our 100th episode. It's an incredible honor to be recognized among the best in communications and media. Huge thanks to all our amazing guests and listeners who made it possible!
What are operators doing with AI? Microsoft Azure’s latest telecom industry blog post details a number of projects that go beyond chatbots to autonomous systems that interpret objectives and execute complex tasks independently. Notables include AI-driven fraud-detection systems for improved network security, generative AI-powered voice solutions to reduce spam calls, and an unnamed North American operator who reduced network CapEx by 10% by adopting an autonomous optimization agent. Sounds super cool!
Telcos are facing a big AI data dilemma. They need diverse, clean, accessible data for effective AI. But all their data is trapped in disparate systems and spreadsheets. Apparently, there was a lot of talk about it at FutureNet World London, the industry event for network management execs. I guess they should have invited me, or someone from Totogi, because BSS Magic solves this problem. A key feature is that it ingests the data from ALL your systems, no matter what vendor, and acts as a universal translator to breathe new life into your applications, helping them work as one. Think of it as a telco specific MCP server. If you’re lamenting the state of your data, BSS Magic is definitely something you want to check out.
Another approach people ask me about is moving data to another system, like Snowflake or Cloudera, and creating AI applications on top of that. I don’t like this idea because it requires the unnecessary (and hard) step of moving your data and THEN building AI applications. This approach is too hard and takes too long to get results. Again, BSS Magic nails it, connecting to the systems you already have in place and getting to work immediately. Don’t believe me? Then give it a try! Telcos can get a free BSS AI agent with Totogi!
AI is cutting costs and boosting efficiency for Globe Telecom. The Philippines-based operator’s GenAI Quality Audit tool has cut annual auditing expenses from tens of thousands of US dollars per month to about $35 bucks, contributing to a 4% year-over-year decrease in total operating expenses. And it’s not stopping there. Globe also plans to use AI to develop hyper-personalized subscriber services—which you can do, too, through Totogi’s PlanAI!
Anthropic has launched a new API for AI-powered search. The tool will give the Claude AI models the ability to search the web, allowing developers to build applications that deliver current information without managing their own search infrastructure. Compatible with Claude 3.7 Sonnet, Claude 3.5 Sonnet, and Claude 3.5 Haiku models, the service is priced at $10 per 1,000 searches and extends to Claude Code, enhancing its ability to access current programming documentation and technical resources. With search being built into the models, are we a step away from the large language models (LLMs) just providing search themselves? If I were Perplexity.ai I’d feel threatened right about now…
On a related note, OpenAI added shopping to ChatGPT. Now, when anyone uses the LLM to search the web, they can get product recommendations complete with images, reviews, and links to merchant websites. The results in ChatGPT are currently organic search results, not paid placements like you see on Google. Plus, ChatGPT is making the recommendations based on product reviews, editorial sources, and user forums like Reddit. Watch out, Google!
The AI space is brutal right now. One day, your AI startup might be valued at $14 billion, and the next day a bigger fish launches a feature that kills your valuation overnight. Net-net: no one is safe, not even Google (which, even two years ago people would have called me crazy for saying). You know what is safe? Mobile networks. No one’s coming for that. So while you have the chance, reinvent your business model, adopt new tech like public cloud, AI, and of course, Totogi—and reshape your P&L for profitability.
Still don’t think LLMs are a commodity? How about when a new LLM model drops that’s 2% better, and it’s super easy to swap out models and use the new one with one line of code? Imagine if you could do that with BSS systems? Dare to dream. Still, some telcos have spent a hundred million dollars betting big on ONE LLM. I said it before: the value is not in the model. It’s in the applications you build. Be sure to bet on the right layer.
Another skirmish in the public cloud vs on-prem debate: Remember back in 2022 when David Heinemeier Hansson (aka @DHH), co-owner and CTO of 37Signals, proudly declared to the world his email service was leaving AWS (all they do is “rent servers”) and was moving everything back on-premises? That repatriation project is almost done, and DHH has spent the intervening time saving only about $1.3 million a year. Meanwhile, AI has exploded, and Zack Kanter called him out on it, pointing out the HUGE opportunity cost with regard to AI. DHH fired back at Zack and obviously disagrees. You know whose side I’m taking. In this debate, DHH is wrong, and Zack is right. Bet on the public cloud every day, and free up your teams to focus on things that move the needle for your telco, like AI.
