Mark Prudhoe, SaaS Business Development Manager, Amdocs ConnectX always brings something new and exciting to the stage at MVNO Nation events.
At MVNO Nation USA 2025, Mark explored the transformational opportunity for MVNOs using agentic AI to automate complex processes and improve customer service.
The difference between agentic AI and “normal” AI
Before exploring what agentic AI can do for MVNOs, we must first make the important distinction between two different types of AI:
- “Normal” AI-enhanced workflows
- Agentic AI.
1. An AI-enhanced workflow
This is a more simple, modular way of using AI, working on singular tasks, like the traditional ChatGPT model.
This typically works by allowing a customer to input a prompt or a query, and the AI tool gives an answer or an outcome. This is quite narrow in its context.
For MVNO customers, this would likely be a chat-based interaction, for example, or receiving some kind of QR code where they can get information on their plan or bill.
This is the more basic, and easier, type of AI to implement for MVNOs.
2. Agentic AI
Agentic AI is a system of autonomous AI-powered agents, each with their own specialism, working both individually and together to complete more complex tasks.
These agents have a shared knowledge base, and they’re orchestrated to collaborate in executing complex processes across multiple disciplines. For example, if a customer makes a complaint about their bill, the agent can take care of that entirely.
Using AI agents in a customer-facing role effectively must start with helping them understand what each user wants, and how to give that to them.
For instance, if a customer who’s going abroad wants an eSIM, the AI agents can do that for them, but that is a multi-step process. So, in that case, the customer would call several agents into action to work together across different areas like orders, payments, and so on.
That collaboration is what makes agentic AI different for MVNOs. It’s a more advanced system, with more advanced capabilities than your typical generative AI tools.
Best practices with agentic AI
For MVNOs looking to use agentic AI to streamline and automate processes, it’s crucial to take a best practice approach to implementation from the very beginning.
Automating inefficient systems, or relinquishing too much control to AI, will result in more challenges in the long-run.
For example:
- Enable the AI agents to work autonomously and independently, but don’t underestimate the importance of human supervision and quality control.
- Integrate existing systems and tools with AI to maximise its capabilities across the business.
- Educate the AI agents, and give them full context, so they understand how to do their jobs properly.
- Allow the AI agents to be continuously learning and self-improving based on each task.
- Don’t overlook security and governance with your AI systems.
An MVNO will typically use a range of multiple agents, which can each be organised into verticals.
Each of these agents will be trained with their on specific telco skills, such as:
- Customer care agents,
- Sales agents,
- Marketing agents,
- Network agents,
- Home agents
- And so on.
And each agent will have its own set of sub-agents, each with their own skills, where an agent resolving calling issues will have sub-agents trained on usage, login, allowances, enrolment, and more.
Underneath all that is orchestration of complex processes and workflows, providing you with work that’s almost entirely autonomous.
The areas in which agentic AI will be best used are processes and workflows that are based on a repeatable series of steps, or tasks with limited possible responses and outcomes.
Don’t let agentic AI compromise your brand or customer experience
Of course, all successful MVNOs are fundamentally built on brands that customers build a connection with and stay loyal to.
It’s important to keep your brand and your customers’ experience top-of-mind when considering technology like this. No new technology should be implemented at the expense of your brand or your customers’ best interests.
If your customers love you because of the personal, human feel of your customer service team, overhauling that with robots may end up harming your customer retention.
Agentic AI can be trained on your brand style guidelines, and can attempt to provide a similar experience to customers. For instance, you can create an AI-generated “avatar” that talks to your customers like a person, trained on brand voice and tone. However, you should still think carefully about whether that’s the experience your customers actually want, and consider where agentic AI is best used and where your brand would benefit from remaining true to itself.
A practical example of agentic AI supporting an MVNO
A perfect example of this was given by Mark, as he explained:
“For a small or medium-sized MVNO one night at midnight, thanks to a recent successful marketing campaign, you’ve received 15,000 new orders that need to be shipped out. Traditionally, that’s done by people in a warehouse. At midnight, there’s only a small team working, and then a big storm hits the warehouse.”
“This could be a disaster, but agentic AI can be working for you round-the-clock to prevent situations where important processes are at risk of being interrupted. With agentic AI, you have an automated system that’s always checking the weather, checking what customer orders need to go out by when, planning any need to re-route shipping, redirect packages, and more.
In a disaster, the agentic AI will know which customers have been impacted, notify those customers that there’s going to be a delay, and answer their complaints, all before your team has even woken up in the morning. That’s the opportunity that agentic AI will bring.”
A reality check on ROI and hidden costs with AI
Mark also pointed out that there are hidden costs to AI that many businesses are overlooking. MVNOs must take a reality check before believing all the hype and blindly adopting new technology just because it’s trendy.
More importantly, businesses shouldn’t be looking to cut costs by removing large sections of their workforce completely. Autonomous AI agents shouldn’t simply replace people, as their implementation creates new human tasks, new responsibilities, and new risks.
Mark said, “I want to make sure we have a balanced argument here. The myth that firms will claim is they’ll be able to cut 22% of your workforce and save you money, but that’s not the reality.”
“The reality in the data we’re seeing shows us you can cut staff costs, but instead of losing people, it’s about changing those roles. It’s about making sure that you have people in your organisation who are capable of using AI.”
By freeing up staff with agentic AI working autonomously in low-value or monotonous workflows, your people are able to do more in higher value areas. This allows more care to be taken in handling your backlogs, improving products, supporting growth and working with customers, and more.
Collaboration and education among the community is needed
Agentic AI is constantly evolving, and its capabilities are advancing at a rapid pace. Even today’s experts will be behind in a few months.
So, the best way for MVNOs to take advantage of these opportunities, while managing the challenges and mitigating the risks, is by coming together as a community.
Sharing knowledge, experience, and ideas with AI will be an important success factor over the coming years.
Mark said, “We need to start building organisational and industry-wide AI fluency. It’s as much about developing people, first, and developing policies around AI.”
The MVNO community must be working hard to ensure we can implement AI in a responsible, ethical, and empathetic way.
Mark added, “It’s ok to be scared of the change AI will bring. To remove that fear, we need to have open, collaborative conversations as a community, to better understand it. And what it’s all about really is using AI to amplify our potential and give a better experience back to our customers.”
Key take-aways and guidance for MVNOs
To successfully implement agentic AI, MVNOs should:
- Understand the difference between agentic AI and generative AI
- Take a responsible, ethical, human-centric approach to AI
- Prioritise what your customers want and stay true to your brand
- Automate repeatable processes to free up your employees for higher value work
- Recognise the hidden costs, and don’t listen to the hype that over-estimates ROI
- Follow best practices, and look to the six principles of agentic AI for guidance
Collaborate with the rest of the community to share knowledge and experience.
