Wade4Wireless

AI in Telecom: 2024


I want to bring up the role that AI will and is playing in telecom. It can be a game-changer but there are a ton of obstacles out there. 

First, I want to point out that Klarna, a “buy now pay later” credit company is already publicly admitting that they have AI assistants doing the job of about 700 people, which incidentally, it laid off about that number in 2023. I know you may find this hard to believe, but here are the articles. When you Google stuff like this, it’s hard to find, that’s why I want to include 2 links for proof. 

Above I am talking about replacing customer service personnel with a chatbot and maybe something a little better. This also entails having the existing people use ChatGPT to research, prepare, and proof anything they do. 

Telecoms already have much of this in place, but I am not sure how much money they really save. Probably fewer inside customer service and marketing people. When it comes to larger telecom companies, like AT&T, Verizon, and T-Mobile, think how complicated their plans and fees are. Can you imagine AI trying to figure them out? It may figure it out, but it seems like they’re constantly changing. In my head I imagine even ChatGPT pushing back and telling them they’re not making sense. OK, that’s in my head. They will bend ChatGPT until even that AI thinks it’s normal. 

Now, for the MVNOs like Visible and Mint where the plans are so simple and flat rate, it makes sense to have automated interfaces. That is why most of these businesses don’t need brick and mortar stores because their plans are simple and straight-forward. Send me the SIM or scan to use eSIM, then get online, open the app, and I’m done once the activation goes through. 

That’s probably why they have smaller legal and marketing departments. With Visible, you get what you pay for and no additional fees. Easy peasy.

Why can’t Verizon do that? Is it the telecom mentality? Does big telecom know better than us mere peons? For those of you that don’t know, a peon (pronounced PEE-on) is a menial worker, referred to as simple or unskilled. That’s me, right? I’m not big and powerful nor am I an executive. I am me.  

However, I do see MVNO and disruptors embracing AI to move past where we are today. Not just to replace people, although that will come, but to:

  • Break into new markets, 
  • Automate connections, 
  • Improve billing and customer service, 
  • Increase uptime and quality of service, 
  • Automate NOC services,
  • Lower network latency and increase throughput,
  • Automate FAQ to a chatbot that actually can talk to users,
  • And probably a hundred things I am overlooking right now. 

Let’s break away from the simple obvious ChatGPT solutions. 

Generative AI Network Solutions

I am not just talking about replacing customer service here. Generative AI is already improving networks for telecom according to Microsoft. They list a bunch of ways companies save but I would like to specifically look at network routing.In this case, the monitoring of the network will go to the next level. While it can’t solve all the problems, what it can do is look at the following issues. 

  1. Traffic improvement and rerouting as required. This required the AI NOC, (I just made that up but pretty cool term, right?), to monitor and analyze the network for anomalies, overloading, underutilization, and other network traffic issues. It could use a combination of real time traffic stats and a predictive model based on the history of the network and that particular router that would be moth a specific route and macro routing view. Think of when your maps program is guiding you and warns you of an accident ahead or heavy traffic and offers to reroute you on a less traveled road. Same concept. If the AI NOC is looking at this and then starts to divert traffic through a less used route to avoid bottlenecks, it would be a huge asset to the network. This has been done through machine learning but now AI looks for anomalies and can make this change on the fly automatically and efficiently without human intervention. This is a step up from circuit switched fallback in telecom. It would be automated and perhaps predictive. 
  2. Failure prediction would be a huge thing. When we could predict failures in the network, we can schedule outages, in theory at least. However, we have to look beyond the network. This isn’t done just by monitoring traffic. It would be to monitor the temperature of the equipment along with version numbers and historic failure rates. That’s why using
    Generative AI would be a game changer because it looks beyond one specific thing to piece together the model for future issues with that piece of equipment in that data room utilized in that way. Think about doing this with thousands of pieces of equipment, all feed your AI NOC. 
  3. Planning and upgrades before the traffic gets out of control, I know we have the surprise events but if AI NOC notices that traffic at point A is increasing over time and incrementally, it can order upgrades and additional fiber months in advance instead of after the fact or days in advance. For those of you that have done this, it’s never a quick process and what appears like a simple upgrade on paper can be a coordination and implementation nightmare. It can rarely get done quickly.
  4. Quality improvements could be monitored by the AI NOC and could look for QoS (Quality of Service) issues in real time. As it “sees” the quality parameters drop on the reporting it can make adjustments on the fly. In telecom networks this is an issue. Just try to use your device during rush hour or at a packed football game, it’s not the best quality of anything on your device. It could simultaneously look at video, voice, and data quality along with latency. This is a powerful tool that, in theory, could look at all these KPIs (Key Performance Indicators) simultaneously. Another thing that AI NOC could do once it’s ready.
  5. Energy improvement could go to a new level. If we could scale back the use of routers, servers, and other equipment to save energy, it would be awesome. In radios they’re starting to do that during slow periods so that those energy hog radios become slightly more efficient. While this is not the best way to do things, it can’t hurt. Also, we could compare equipment in real time instead of relying on the data sheet to tell us what it should do. If we could monitor the energy usage in real time and provide feedback we could go back to the OEM and complain instead of waiting for 3 months of utility bills to stack up. 
  6. Project management is another solution that will be somewhat automated. Think about the scheduling and planning of the deployments, the collection of documentation, and so on that could be managed by fewer people because the AI PM Bot is always on the case to make sure all tasks are being updated and completed without all the complicated spreadsheets we use today.
  7. NOCs can be local. The US can have mainland NOCs once again using AI NOC because the offshoring being used today, mostly in India and Mexico, will be brought stateside like it was 20 years ago. 
  8. AI works 24/7 without overtime, holiday pay, or complaining. While there are still downsides, this is a huge incentive.

We have to go beyond the network and into the pre-deployment and deployment phases of the network build and expansion.

AI for network Deployment

From my experience, I will tell you companies are constantly looking for automation and improvement to replace high paid American workers. That is, until it comes to the executive’s jobs, that is where they draw the line and preach leadership is needed, not realizing the real problem is staring at them in the mirror every morning. 

  1. Site Acquisition is something that is already trying AI for so many things. However, site selection based on existing fiber routes, specific tower company ownership, and concentrated population or traffic requirements could become a whole lot easier and automated. Nice!
  2. Project management (PM) is being automated but still needs local talent. Telecom companies, at least the front lines, still want someone they can talk to and discuss how to move forward. This has not changed. Behind the initial PM there must be some intelligence. AI could replace this soon, but the customer is uncomfortable with not having a real person to interface with. PM has not been replaced by any apps just yet, but there are apps to help it along. 
  3. Solutioning and planning is one thing they are trying to replace. The one thing holding it back is how scattered all the systems are and experience. AI intends to replace the experience, and it will someday, but so far it has failed miserably. This is about 8 years away. 
  4. Documentation is pretty automated; this is the one thing that is being addressed today. However, did you ever work with a telecom company, any OEM or carrier or even a cloud company? They all have their own way of doing things. They all have a ton of spreadsheets. Some use Smartsheet’s, which is still a type of spreadsheet. So far, the only standard is the spreadsheet, and all of them are unique. They all have a unique sheet to capture data. Do you know why? Because each individual person has their way of looking at it, and then converting it into their company or even departments systems. It’s crazy but in each department, they have their own systems that could be and probably are unique to that department. I see this is many companies and until the people are forced to change, the systems won’t. Remember that each system is a kingdom protected by that person’s group, they create a moat that makes it so hard, (expensive), to change. Adapt or die, but adapt to their system, not something new. 
  5. As for the permitting, they still need CAD drawings that are stamped and approved for the load. This is still an engineering stamp to put the responsibility on the architectural company. This won’t change. What has changed is the digital twin technology which can update the drawings using AI to add the new specifications and equipment to the existing drawings, eliminating simple tasks and is available today. 
  6. Automation in all the above steps is making a difference, lowering some costs and improving our timelines.  The simple tasks are being automated and eliminated today. While this has eliminated many overseas jobs, specifically India (and even Mexico), it has created tech jobs in those same regions. Simple tasks are being eliminated but still farmed offshore for setup and quality checking.
  7. Tracking expenses and spending is being automated today, but the decision making on how to move forward still needs tons of approvals. Not sure what will change.
  8. Procurement is another thing they thought could be replaced. While there is automation to help us find people, we still rely on LinkedIn and word-of-mouth for most of our contractors. We have to vet them and make sure they have all the qualifications. 
  9. Safety tracking and updates are another thing that should be tracked by AI because humans suck at this, at least that has been my experience. My friend Cliff Wilcox started QRedify which takes a QR code that each worker could have and each item could be verified. Pretty cool right? No more being lied to by shady vendors. 
  10. Onsite safety verification through video

I hope that helps show you where we are today and gives you some direction on the future.

Why Telecoms will change at a snail’s pace…..

Because they are set in their ways, they think they know better, and will do what they can to convince you of that. 

They work too hard to get rid of skilled labor while keeping the low paid menial staff in place because most of it is overseas. 

I am going to give you some examples:

  • Who do Telcos lay off? The front-line workers and mid-level managers Not many executives unless stock prices decline drastically. It takes the shareholders to hold them accountable. Think of Crown Castle’s issues with its shareholders in 2024.
  • Did you see the 5G upgrade, it repeated the same style the 2G, then 3G, then 4G upgrades did. Teh carriers paid more for spectrum, then back to business as usual.
  • Do you see a move to Cloud? Eventually but it was slow and painful. It’s still happening.
  • Do you see a change in customer service? When they make changes, it’s slow.

Creating change in a big company is slow, painful, and (more often than not) horribly executed. They publicly say how great it went, but inside it rarley feels that way.

They also have systems that are way too complex and managed poorly by committees. Each goup treats their tool or process like it’s a kindom, and they defend it until the end. In this case, defending it means keeping it funder for as long as possible. As long as it’s funded, those groups all have jobs.

To be honest, I was hoping DISH would set a new standard by bringing in some entrepreneur thinking executives alongside their telecom executives, but I didn’t see that happen. Same old business as usual.

Then I had hoped IBM, Amazon, Google, or Apple or someone would bring something new to the table in telecom. Here are my observations:

  • Apple created the iPhone and all their subscription services. They didn’t care what carrier you were on or even if you were on a carrier’s service, just use their apps and subscribe so they can charge you for memory and iTunes. Facetime was cool and it is still free.
  • Google followed Apple’s model, and added core services. Well, it’s really a telecom cloud core, mostly for IoT, but it’s a step in the right direction. I was hoping it would evolve. Google is pushing for open standards in 5G, which the telecoms wanted to destroy the OEM providers, but now they are backing away because they see new competition from outsiders like Google. 
  • IBM bought Santaca to get into the telecom consulting and integration markets. Good luck.
  • Microsoft Azure made a daring move when it purchased Metaswitch and got into the core business. This will be a real game-changer since AT&T is already planning to offload the 5G core services to Azure. This will make their life easier until Microsoft has too much control. That may happen when Ericsson is in control from the RAN side. (Does anyone think someday AT&T will just be offloaded to other companies? I wonder what FirstNet is thinking?)
  • AWS also has IoT core functionality that is being used by businesses. While they have signed agreements with telecom OEMs and carriers, it would be just as easy for me to sign up my devices in amazon. This is the closest thing to disruption I have seen.
  • I haven’t seen any new disruptors break in to make big changes, yet. I plan to help anyone who can do it. I had high hopes for DISH, but when they made the deal with T-Mobile, they followed the same path as the other carriers. Boring!

So, here is my take on the state of AI in telecom today.

So, what does this mean to the Telecom workforce?

It depends who you talk to. When the telecom companies talk to the government, they will say it’s alleviating the workforce shortage. The reason for the workforce shortage is because they laid off so many people, that’s the reality right. They don’t want to take on the huge workforce again, so this actually solves their problem.

Don’t worry, it’s only a matter of time until they realize AI can also replace executives for decision making, that is going to be the real money saver in companies.  However, I do see AI picking up some unskilled and skilled labor jobs in the near future. Just look at what the tech companies are doing today.

The one thing we have going for us is that these large companies have antiquated processes and tools that are, at best, a kluge or miscellaneous interfaces and tools that the company didn’t want to invest into improvements or someone with power is protecting their outdated and inefficient tool kingdom. I mean, how soon can they possibly change that without an upfront investment? 

Will they make the investments? Sure, when Accenture or McKinsey tell them too. IBM might even make the suggestion from their consulting side of the business. Then big telecom will follow like sheep. 

New book released!

More:

Offers available:

  • All my older Deployment books and SOW training can be bought here!
  • Need unlimited mobile service for a good price? I use Visible and I like them. Think about Visible. Check out this deal https://www.visible.com/get/?ZHfvq, not too shabby.
  • Looking for a new credit card? I have an Amazon Prime card because I get 5% on Amazon purchases along with 2% on restaurants and gas and 1% on everything else. I love it! Use this link to get one.

Small business owners: Looking to sell or merge? I have partners looking to partner with small tech & IT businesses. Details:

  • Looking for companies in IT, Fiber (indoor or outdoor), wireless, Wi-Fi, FWA, Venue or DAS.
  • Looking for owners ready to sell or retire.
  • Smaller companies, maybe 4 to 15 employees.
  • Concentrating on the East coast, but open to US based businesses,
  • Email me at wade@techfecta.com.

Looking for financing to grow, expand, or purchase a business? I have a partner who can help:

  • Who they can help:
    • US based,
    • Someone looking for $10M and up,
    • Needing to grow, consolidate, or merge,
    • Email me at wade@techfecta.com.
  • About the Investor;
    •  Made 155+ completed transactions,
    • $2.4B+ deployed,
    • $1.4B of assets,
    • Focused on Technology, IT, and Wireless,
    • Been around over 15 years.

Leave a comment