We live in hope, according to the old saying, but enterprises’ hopes and expectations aren’t exactly aligning for next year. While all 223 of the enterprises I’ve gotten data from are hoping for real and meaningful technology advances in 2024, only 71 expect that to happen. Technology budgets are not expected to rise; only 58 enterprises said they thought their IT and network budgets would be slightly up, and none of the operators believed that. Cost management is the priority for 217 enterprises and all the operators. According to the enterprises I surveyed, their expectations are for a better 2025, not so much for 2024 -- but that doesn’t mean everyone is taking a year off. They’re looking at some new technologies, yet missing some important things along the way.
The hottest single technology enterprises are looking at for 2024, one that interests 189 of the enterprises and used by 30, is SASE/SD-WAN. Even here, enterprises are missing an important point, which is that the two technologies have almost-competing value propositions, and are really converging into one.
Many enterprises already use SD-WAN as a way of connecting small remote sites to their VPNs. SD-WAN rides on an Internet connection so it’s available anywhere networking is available, where the usual VPN technology, MPLS, may not be. SD-WAN is far less expensive than MPLS VPN technology, too, so starting in 2022, a few enterprises started switching to SD-WAN for all their remote sites.
SASE, or Secure Access Service Edge, is a cloud-level tool designed to serve as the on-ramp from the Internet to the cloud and usually onward to the data center. Since nearly all enterprise data center connections are made via an MPLS VPN, SASE includes SD-WAN features to make the VPN connection. Enterprise adoption of SASE started slow but has been picking up, and 114 of the enterprises on my list expected to consider it in 2024.
What makes these technologies competitive is that almost all the enterprises who plan to consider SASE next year are also looking at using the cloud as the on-ramp to their core data center applications. That would mean that remote office workers could access their applications using the cloud and SASE, which means they wouldn’t need a remote office SD-WAN connection at all. The more SASE, the fewer SD-WAN branch sites.
Obviously, if SD-WAN can be included in SASE, you could run SD-WAN as a process in the cloud, which means you could run it as a process in a remote office site, in a small server. SASE could be run in that same server, so it’s likely that in 2024 we’ll see more and more SD-WAN technology migrating to a software basis rather than as a device, and as that happens, SD-WAN will fuse with SASE. It won’t be an instant transformation, but progress will be relentless, and eventually almost everything will end up SASE.
OK, SD-WAN and SASE are exciting. Is there a technology that doesn’t exist and should? Yes, and it’s 5G slicing for business Internet. Operators have spent a boatload on 5G deployments and reaped almost nothing in terms of incremental revenue. Given the SASE shift just mentioned, and the current SD-WAN interest, it would make sense for network operators to launch Internet services designed for enterprise remote sites. Not only could this improve QoS for business users, addressing the most common objection to the use of SASE/SD-WAN, it could be a camel’s nose for other business services, especially security.
Operators would love to find some way to charge for the capabilities of 5G, the most significant of which is network slicing. A “slice” is a piece of shared infrastructure that separates traffic from the rest. In mobile services, its largest likely application is supporting MVNO providers, retailers who wholesale mobile services from another provider who actually has spectrum and infrastructure.
With the increased interest in fixed wireless access (FWA), enterprises could hope to have enterprise-focused network slices available to them to support remote offices, but none of the enterprises I chatted with told me they were looking at this capability. Where there was a tickle of interest was in the 5G Core element of network slicing, because 5G Core facilities could be shared with traditional Internet access. That would extend better QoS to traditional business Internet services, including fiber and CATV access.
Where would these other service features, like security, come in? Remember that SD-WAN and SASE are converging into a cloud-hosted on-ramp? Well, an SASE microservice could be included in a business Internet service, making the new service a QoS-enhanced VPN service. If it were based on FWA, the most determined cable-seeking backhoe couldn’t eat it. An SASE microservice could provide backup for a true wireline connection, be set up quickly in a temporary area in the event of a natural disaster, or simply to add capacity quickly. For all these capabilities, there would be a price, so operators would see some revenue. Which is why, even though there was little recognition of the opportunity among enterprises, 48 of 88 operators were very interested.
How about collaboration? Of 244 enterprises in my survey, 227 said that they were very interested in new collaboration technology for 2024, and the major technology interest they cited was AI, but only 11 made any connection between the two. They missed something important, because one of the missions enterprises see for AI is customer support, and it’s a very small step from there to our last hot technology—collaborating with AI instead of with other workers.
A couple years ago I did a survey of 277 users on interpersonal communications in a work environment. We tend to think “meetings” in the sense of a meeting room with a white board as the collaborative center, but in my survey, 88% of all interpersonal communications reported and 92% of the total time communicating involved only two people, and took place at the desk of one or the other. Even if we limit the field to face-to-face collaboration, people still spent five times as many hours in two-party discussions than they did in F2F meetings.
What’s going on in these sessions? Help, mentoringwork reviews, you name it. What’s interesting is that almost every collaborative task was also on the list of things that enterprises hoped an AI chatbot could offer in a customer support role. So AI chats are good enough to depend on in customer relations but not good enough to field some of these two-party worker interactions? Bull. A few enterprises even noted that some of the same topics were likely to come up in both places, so how much worker time could be saved by having these topics handled with an AI chat instead of a chat with a human?
Even some complex tasks normally handled through two-party communication could be handled with an AI chat. Rather than focusing AI on doing the complicated stuff, ask it to review human output, which is a much lower apple to pick. Software developers do code reviews, and AI can already do much if not all of that. Proofreading and text or presentation tune-ups are obviously an AI opportunity, and some spreadsheet work can already be validated using AI. Facts in a press release, details in product brochures or even manuals, can be validated via AI rather than through human interaction. Enterprises already love AI for its potential, but they’re not even close to realizing what might be the easiest applications to launch, and the most valuable.
Time to summarize. We’re in the most revolutionary period in technology for decades, maybe for all time, with the reality of the Internet combining with the promise of AI. You don’t justify a revolution with trivia, you have to think big, think not only out of the box but ahead of it. That’s what enterprise technology planners, planners like you, need to be doing in the new year. Have fun, because this is the start of a wild ride to the future.
