A major Southeast Asian telco reduced configure-price-quote (CPQ) order time by 80% with AI agents. The old way took over 5 minutes and 50+ clicks to create a single CPQ order—and that’s the expert’s fastest time. Now, users simply type a natural language prompt. The agent builds the full, validated order in seconds, executes it in CloudSense, so the enterprise sales team can focus on what it does best: selling. This is what AI-powered CPQ looks like. Want this for your enterprise sales team? Give me a call! 📱
The large language model (LLM) wars continue unabated, with Anthropic releasing Claude 4.1 just days before OpenAI launched GPT-5. What are we fighting over? Claude’s moat of coding superiority. GPT-5 reportedly outperforms Claude 4.1 on multiple tests, including SWE-bench Verified (a test of real-world coding tasks), scoring a slightly higher 74.9% versus 74.5% for Claude Opus 4.1. GPT-5 features major improvements in accuracy (fewer hallucinations), reasoning, user experience, and reliability. Plus, it’s free for everyone to use. But your actual mileage may vary according to real-world tests done by users on X.com and Totogi’s own experience. So, what do you think? Which has the best model for coding as of the end of August?
OpenAI just dropped two open-weight reasoning models you can run locally. The result of billions in R&D, the models are completely free, private (no data leaves your device), and can operate offline. Even better, they’ll personalize the results they give based on individual users’ usage patterns. The “open weights” part means you can modify the models themselves, fine-tuning them for specific purposes. With the enhanced privacy, this new model makes it possible for regulated industries to use AI. Here’s a great one-pager about the release from Greg Isenberg. With these models being small enough to put on a smartphone, it’s starting to open up some exciting telco use cases. Totogi’s already trying it out! Give me a ring ☎️ if you want to learn how.
Delta Air Lines is rolling out AI-powered dynamic pricing that customizes ticket prices to passengers based on their individual willingness to pay. The personalized pricing tech acts as a “super analyst” to set real-time custom prices based on customer behavior. Telcos can do something similar with custom plans and services using Totogi’s killer Plan AI—a proven way to grow revenue with hyperpersonalization. Give it a spin for yourself and see!
You can’t rely on old-school vendors for your AI strategy. They don’t get it, and they move too slowly. Model Context Protocol (MCP), Anthropic's open standard for connecting AI models to data sources, is being driven by AI labs and cloud platforms—while Nokia and Ericsson are still “tinkering in labs,” according to analyst Dean Bubley. We’ve seen it before with 5G slicing, cloud-native cores, and WebRTC, where vendors hoped to monetize the innovation but moved at a glacial pace. MCP could become as fundamental to AI workloads as TCP/IP is to data networks. I agree MCP is pretty critical to use, but it's not the whole enchilada. Totogi uses MCP in BSS Magic to connect to legacy applications, access their data, and add AI—but you're still missing a critical component: the telco ontology. Shoot me a note and our team will set up a demo to show you how to really build scalable AI for your telco.
Rakuten and Optus just joined the agentic AI party, but they’re making the classic enterprise mistake: building AI agents with one-size-fits-all tools that don’t understand telco. Here’s the problem: generic AI agents struggle to scale across telecom operations because they lack industry context. If your agent can’t distinguish between provisioning errors and billing disputes, it’s just an expensive toy, not an enterprise-grade business tool. Seeing the disconnect, Gartner predicts 40% of agentic AI projects will collapse by 2027. Smart operators are using telco-specific platforms like BSS Magic that has a telco ontology built-in that can scale up meaningful automation across billing, provisioning, and customer management, delivering results quickly.
Juniper Research forecasts network API revenue will exceed $8 billion by 2030, exploding from $284 million in 2025. The report identifies Know Your Customer (KYC) APIs as the next big revenue driver, offering stronger identity assurance by verifying users against operator data including subscriber info, SIM, and device details. Early wins like SIM Swap and Number Verification APIs are also building confidence in the ecosystem. Finally, telcos are waking up to what cloud-native operators have known all along: your network isn’t just infrastructure—it’s a platform. Operators can become the identity backbone for everything from e-commerce to gambling, charging for bulletproof verification that only they can deliver. The question is, can telcos get enterprise developers to put network API calls into their code? We shall see…
IBM just clutched its pearls about 63% of companies having “no GenAI governance policies.” My reaction? Um, maybe that’s OK? I think it’s a good thing these companies aren’t wasting months building governance frameworks before knowing how generative AI (GenAI) will be used in their business. The fastest way to kill GenAI momentum is to drown it in committee meetings. Instead, start small and prove value first. Pick a low-risk use case, get some quick wins, then layer in controls. The operators moving fast on practical GenAI will lap the ones still forming governance committees. Net-net: stop overthinking and start using.
Parents in Maine are ditching smartphones for landlines to fight screen addiction—and honestly, it’s kind of genius. A South Portland pod of 20 families installed old-school corded phones and plot twist: the kids love it and are becoming better conversationalists without endless notifications. With phones being removed from schools, too, and studies showing that kids shouldn’t get mobile phones until age 13, will this next generation escape cellphone addiction and swap it for an AI habit? 🤔
