You don’t have to upgrade to get AI
The core problem: Legacy vendors like Amdocs sell agentic AI bundled with multi-year systems-integration contracts because their revenue depends on billable headcount. They say you have to upgrade your BSS before the AI arrives. You don’t. The Totogi Ontology runs on the systems you have today. Agents reach production in days, and you decide later whether to upgrade anything.
On November 3, Amdocs had a terrible, horrible, no good, very bad day. It told investors that fiscal 2027 revenue could fall as much as 18%, because T-Mobile, its second-largest customer, has cut spending and turned a dispute over recent agreements into a legal fight.
It also announced what looked like good news: a shiny new AI deal. Vodafone Business UK named Amdocs the Lead Systems Integrator for the agentic transformation of its B2B business. The release spelled out the order of work. First, Vodafone’s legacy BSS will be consolidated onto Amdocs’ platform. Then the AI will arrive, with Amdocs’ people running the project throughout.In short, Vodafone is buying an upgrade to get AI, and paying for Amdocs’ army to install it.
The stock fell 40% the next morning.
Every incumbent in your estate is re-labeling its services business as agentic AI, and the pitch is the same every time: a platform, a transformation program, and their people to make the AI work. Amdocs is only the most visible example. Strip the AI off the cover and you’re looking at the same multi-year integration contract you’ve signed before.
You shouldn’t have to upgrade to get AI. Nor should you have to hire an army of consultants.
The estate forced you to hire the army
If you’re a telco CEO or CIO who has signed tens of millions of dollars in systems integration contracts over the past decade, none of this is on you. You inherited billing, CRM, rating, and provisioning systems built across three decades by vendors who had every reason to keep them separate. No two of those systems agree on what a subscriber or an order is. To launch a tariff or consolidate a ledger, you had two choices: hire people to translate between the databases by hand, or replace the databases until they agreed. The headcount was a rational response to the architecture you inherited.
Now your board wants agentic AI in production, and you can’t build it yourself because the talent math doesn’t work. You cannot outbid OpenAI, Anthropic, or Meta for frontier AI engineers. So when your incumbent vendor offers an “Agentic Operating System” with the people to wire it up, signing that contract feels like the only practical way forward.
Don’t do it.
What Vodafone is actually buying
For years, every executive in this industry has followed the same sequence: first consolidate, then modernize. Once the estate is clean, you earn the right to do something intelligent with it.
Under that assumption, Vodafone’s decision looks rational. The problem is the assumption. The first step, the upgrade, is what takes years, costs hundreds of millions, and puts an integrator in the middle of your business—and all of it happens before a single AI agent has delivered a thing.
In the second step of the sequence, every rule your people explain and every exception the integrators resolve gets encoded inside the vendor’s platform, by the vendor’s people, under the vendor’s contract. By renewal, the conversation is no longer about a software fee. You are negotiating for access to the logic underpinning your own business. Just ask T-Mobile how that conversation goes.
Vendors sell this sequence because it serves them. In fiscal 2026, managed services produced $3.02 billion of Amdocs’ $4.7 billion in revenue, 64% of the company. For a business that bills by the human hour, software that works on its own is an extinction event. The people are the product. The agents are the brochure.
So, the test is simple. If the agentic AI needs an upgrade and a systems integration before it can start, you’re really buying a transformation with an AI logo on it.
In the AI age, reverse the sequence
An ontology resolves the disagreement between your systems in software. The Totogi Ontology is built on the TM Forum SID model that most of your systems already map to. It sits on top of your existing BSS and OSS, deployed in weeks, with nothing ripped out and nothing migrated. It gives you one live model of your operation, with every subscriber, rating rule, network element, and permission reconciled across the systems you already run. The ontology handles the semantics and the AI works on the mess. You no longer have to clean it up first, and you no longer have to upgrade.
An operator with an estate like Vodafone Business’s could have its first agent running on today’s systems. Later, once the agents have shown where the fragmentation actually costs money, it could decide whether any consolidation is worth it.
The talent equation changes with an ontology, too. You won’t need a large internal AI department or an offshore body shop. A small Totogi team works alongside your best operational leaders, and encodes the expertise in their heads into the ontology once. The ontology defines what each agent is allowed to do, so an agent cannot attempt an action your business hasn’t authorized. Your agents then carry out that judgment millions of times, the same way every time. When our engineers leave, the agents keep running, because the learning lives in rules you own rather than in a consultant’s head. You no longer need to pay someone to tell you how your business runs.
When software replaces manual translation, results show up in days rather than fiscal years:
- Network CapEx runs on revenue instead of assumptions. VIVA Bolivia connected its commercial and network systems, vendor and homegrown, on-premises and in Azure, into one model on the Totogi Ontology without replacing any of them. Its CEO now sees which sites earn their keep and which never will before he spends on a tower. Network investment is the first decision the model improves, with more behind it.
- Zain Sudan started using the Totogi Ontology with its existing systems and existing data. Dormant cell resolution went from 48 hours to 30 minutes. No system was replaced. Decisions got faster.
- Every action compounds into rules you own. Every order created in 50 seconds instead of five minutes and every retention offer executed leaves a decision trace in your ontology that makes the next agent smarter.
The seat at the head of the table is open
For three decades, the largest software company in our industry has operated as a human services organization wrapped around a database. The operators working with Totogi have already seen the other side: agents in production in days, decisions made on live data instead of averages, and a handful of engineers with agentic loops running around the clock doing the work that used to take an army. That is what AI was supposed to feel like.
Stop funding your vendor’s outdated business model. If you have a mandate to put agentic AI into production this year, bring in Totogi. We’ll put the Totogi Ontology on the stack you run today in weeks and put our engineers on your hardest operational problem. Give us fourteen days on a live use case and I’ll show you how fast your telco moves when the software actually does the work. Reach out to me directly or book time with my team.
Recent Posts

Get my FREE insider newsletter, delivered every two weeks, with curated content to help telco execs across the globe move to the public cloud.
Get started
Contact Totogi today to start your cloud and AI journey and achieve up to 80% lower TCO and 20% higher ARPU.
Explore
Telco AI needs an ontology. Should operators buy one and customize it, or build their own from scratch? Here’s how to decide.
Book a demo
Let us show you how the Totogi platform can unify and supercharge your legacy systems, enabling AI to understand, reason, and act across your entire estate.
Imagine
Every telco has a legacy system it can’t escape. Hear from the CTIO at Denmark’s biggest telecom provider about how AI helps him slay the “dragon in the basement.”
Frequently Asked Questions
No. Legacy vendors push “consolidate first, modernize later” because the upgrade is what they’re really selling: multi-year systems integration contracts billed by the human hour. The Totogi Ontology runs on top of the BSS and OSS you already have, with nothing ripped out and nothing migrated. Agents can reach production in weeks. Once they’ve shown you where fragmentation actually costs money, you can decide whether any upgrade is worth doing.
Amdocs is the Lead Systems Integrator on Vodafone’s B2B “agentic transformation,” and the deal requires Vodafone to first consolidate its legacy BSS onto Amdocs’ platform. If agentic AI needs a multi-year integration project before it can do anything, the software can’t do the work and humans are still doing the translation. Managed services made up $3.02 billion of Amdocs’ $4.7 billion in fiscal 2026 revenue, or 64%. Software that works on its own threatens that model. So the consolidation is the product and the agents are the brochure.
A telco ontology resolves the disagreements between your systems in software instead of with headcount. The Totogi Ontology is built on the TM Forum SID model, which most telco systems already map to. It gives you one live model of your operation, with every subscriber, rating rule, network element, and permission reconciled across billing, CRM, rating, and provisioning. The upgrade was always a semantic project disguised as a software project. Once software handles the semantics, AI can work on the mess without you cleaning it up first.
In days, not fiscal years. The ontology deploys on your existing stack in weeks. Zain Sudan cut dormant cell resolution from 48 hours to 30 minutes using its existing systems and data. VIVA Bolivia connected its commercial and network systems into one model without replacing any of them, so its CEO can see which sites earn their keep before spending on a tower. Give Totogi 14 days on a live use case, and we’ll prove it.
No. You can’t outbid OpenAI, Anthropic, or Meta for frontier AI engineers, and you don’t need to. With the Totogi Ontology, a small Totogi team works alongside your best operational leaders and encodes their expertise into the ontology once. The ontology defines what each agent is allowed to do, so agents can’t take actions your business hasn’t authorized. The ontology belongs to you, not to a consultant, so the agents keep running after the engineers leave and you won’t be negotiating for access to your own business logic at renewal.