- AI reshaping telecom from the core
- POTS elimination: the clock is ticking
- 5G business use cases finally maturing
- SASE market consolidation
- The accelerating decline of MPLS
- UCaaS becoming commoditized
- Carrier consolidation and what it means for buyers
- What IT leaders should do in 2026
- Frequently asked questions
If you've been in IT or operations long enough, you know the feeling: a trend gets declared, vendors start evangelizing it, analysts publish reports, and then—sometimes years later—something actually changes. Telecom in 2026 is not that story. The shifts happening right now are structural, carrier-driven, and in several cases, irreversible.
POTS lines are being turned off. MPLS contracts are expiring and not being renewed. The SD-WAN and SASE markets are consolidating. AI is being woven into every layer of call routing, network management, and billing. And 5G, after years of hype, is beginning to deliver real value for specific enterprise use cases—even as it remains oversold for others.
This guide is not a technology forecast. It's a practical survey of what's actually changing in business telecom right now, what the implications are for IT leaders and procurement teams, and what specific actions are worth taking in the next quarter and the next year. We'll call out the hype where it exists, and point to real decisions where real decisions are needed.
AI Is Reshaping Telecom from the Core
Artificial intelligence has been a buzzword in telecom for years, but 2025–2026 marks the point where it's moving from marketing language to operational reality. The change is showing up in three meaningful places: customer-facing voice systems, network management, and billing and security analytics.
Conversational AI in call centers and IVR. The traditional touch-tone IVR ("press 1 for billing, press 2 for support") is being replaced at scale by conversational AI systems that understand natural language, handle complex multi-step queries, and escalate to live agents when needed. Vendors like Google CCAI, Amazon Connect, Genesys, and Five9 have deployed large language model-powered IVR that can resolve a significant percentage of inbound calls without human involvement. This isn't just about cost reduction—the customer experience is measurably better when a caller can say "I'm calling because my last invoice was $400 higher than usual and I need to understand why" and get a useful response.
Agent assist and real-time coaching. For the calls that do go to a live agent, AI tools now surface relevant information in real time—pulling account history, suggesting responses, flagging compliance risks on regulated calls, and transcribing and summarizing the conversation for post-call review. RingCentral, Zoom, Microsoft Teams Phone, and most enterprise UCaaS providers now include some form of AI assist either natively or through integrations. The question is no longer "does this exist" but "do you have the configuration and training discipline to make it work."
AI-driven network optimization in SD-WAN. The major SD-WAN platforms—Cisco Viptela, VMware VeloCloud, Fortinet, Cato Networks—are all applying machine learning to traffic routing, anomaly detection, and performance tuning. Instead of static QoS rules, these systems dynamically shift traffic based on real-time circuit performance data, application behavior patterns, and historical baselines. Practically, this means fewer manual tickets for "voice quality is bad" and fewer emergency calls to your WAN vendor when a circuit starts degrading.
Predictive maintenance and anomaly detection for billing. AI is also being applied to telecom expense management (TEM), where it can flag billing anomalies—unexpected charges, rate changes, usage spikes—before they appear on invoices as line items you've already paid. Several TEM platforms now use machine learning models trained on carrier billing data to spot errors that human reviewers routinely miss. Given that billing errors on large enterprise telecom accounts often run into thousands of dollars per month, this is a real return on investment.
If you have an inbound call center or a multi-level IVR, request a demo from your UCaaS provider's AI team. Specifically ask what your current call containment rate is and what their AI IVR can realistically deliver. The gap between those numbers is your opportunity—and your competitors are likely already pursuing it.
POTS Elimination: The Clock Is Ticking
Plain Old Telephone Service—the copper-wire analog phone lines that have existed since the 1870s—is being formally retired across the United States. This is not speculation; it is a proceeding already in progress under FCC Order 19-72, which gave carriers the authority to petition for copper network retirement on a state-by-state and region-by-region basis.
AT&T, Verizon, Lumen, and Frontier have all filed active petitions. Approvals have been granted in multiple states with retirement timelines ranging from 24 to 36 months from approval. In many markets, those timelines are already counting down. Businesses that have not started POTS inventory and replacement planning are falling behind.
Who is most exposed. The businesses with the greatest risk are those using POTS for something other than voice calls: alarm system dialers that call out to monitoring centers, elevator emergency phones required by building code, legacy fax machines on dedicated lines, building automation and HVAC controls that dial out for alerts, and medical devices that use POTS as a failsafe. These systems cannot simply be "moved to VoIP"—they require either purpose-built analog-to-IP gateways (sometimes called POTS-in-a-box devices from Ooma, Lumen, or Bandwidth), replacement hardware with native SIP support, or cellular-based alternatives.
What the transition actually looks like. Carriers will issue a notice—typically 24 months in advance—specifying when service will be discontinued. They will usually offer their own replacement service, which is often expensive and poorly matched to the actual use case. The businesses that handle this transition well are the ones that start the inventory process now, understand what each line is doing, and match each line to the right replacement technology before the notice arrives.
For voice lines: SIP trunking or UCaaS. For alarm dialers: cellular backup devices or POTS replacement gateways. For elevator phones: consult your elevator maintenance vendor—they have approved solutions. For fax: cloud fax services (eFax, RightFax) cost a fraction of a POTS line. For building controls: cellular IoT or modern IP-based monitoring devices.
Pull your telecom invoice and count every POTS line. For each one, document what system or device it connects to. That inventory—which most businesses don't have—is the entire foundation of a POTS replacement plan. Don't wait until you receive a carrier discontinuation letter to do this work.
5G Business Use Cases Finally Maturing
5G has been overpromised since its commercial launch in 2019. Enterprise IT leaders have been told for years that 5G would replace wireline connectivity, transform manufacturing, enable real-time everything, and make every other connectivity technology obsolete. That hasn't happened—and a clear-eyed assessment of where 5G actually delivers value in 2026 is more useful than continued hype.
What's actually working: private 5G networks. Manufacturing facilities, distribution centers, and large campuses are deploying private 5G networks that operate on licensed CBRS spectrum or mmWave. These environments benefit from 5G's low latency and high device density in ways that Wi-Fi cannot match. A large automotive plant with hundreds of autonomous guided vehicles (AGVs), robotic arms, and sensor arrays can use a private 5G network to coordinate these devices with latency under 10ms and no contention from general office traffic. This is real, it's deployed, and it's delivering ROI in the right environments.
What's working: Fixed Wireless Access (FWA) as enterprise backup. FWA—broadband delivered via 5G radio to a premises receiver—is proving genuinely useful as a redundant WAN path for locations where digging trenches for fiber is impractical or as a fast-to-deploy backup for primary broadband failures. T-Mobile and Verizon are both aggressively selling FWA to enterprises. It's not the right primary connection for a large office, but as an SD-WAN failover path it's a cost-effective alternative to a second fiber circuit or expensive 4G cellular backup.
What's still mostly hype: 5G as primary enterprise WAN. Outside of specific edge cases (temporary locations, mobile operations, remote sites with no fiber access), 5G is not yet displacing dedicated enterprise broadband for primary connectivity. Latency and throughput are adequate for many workloads, but reliability SLAs, traffic prioritization guarantees, and support from carriers are not at the level enterprises require for mission-critical WAN connections. This may change, but it hasn't changed yet.
The practical question for IT leaders is not "should we do 5G" but "which of our specific connectivity challenges would 5G solve better than the alternatives?" That's a site-by-site analysis, not a strategic mandate.
Identify your three hardest connectivity problems—sites with no fiber, backup paths that are inadequate, or campus environments with dense IoT device requirements. Run a 5G feasibility check specifically against those problems. Don't evaluate 5G as a general replacement for your WAN architecture.
SASE Market Consolidation
Secure Access Service Edge (SASE) was coined by Gartner in 2019 and has since become one of the most crowded acronym spaces in enterprise networking. The concept—combining network connectivity (SD-WAN) with network security (ZTNA, SWG, CASB, FWaaS) into a single cloud-delivered service—is sound. The market that emerged around it was messy: dozens of vendors offering various combinations of those components, with most enterprises ending up with stitched-together multi-vendor architectures that were complex to manage and difficult to support.
In 2026, that market is consolidating. The clear leaders—Cato Networks, Palo Alto Prisma SASE, Zscaler, and Fortinet's SASE offering—are pulling ahead of the pack. Each takes a somewhat different architectural approach: Cato built its own global backbone and delivers everything natively; Palo Alto Prisma combines acquired assets into a unified platform; Zscaler built its reputation on cloud proxy security and has been adding SD-WAN capabilities; Fortinet leverages its existing firewall installed base with a cloud-delivered overlay.
What this means for buyers. The key decision is no longer which niche SASE feature you need from a specialist vendor—it's whether you want a full-stack single-vendor architecture or a best-of-breed multi-vendor stack. The total cost of ownership argument increasingly favors full-stack: fewer contracts, one support escalation path, unified dashboards, and no "it's the other vendor's problem" finger-pointing when something breaks.
But full-stack single-vendor SASE is not right for everyone. Enterprises with large existing Cisco, Palo Alto, or Fortinet investments may have compelling reasons to extend those investments rather than replace them. The evaluation framework should start with your actual security architecture, your SD-WAN maturity, and the gap between what you have today and what a zero-trust access model requires.
If you're still running legacy VPN for remote access alongside a separate SD-WAN deployment and a third-party next-gen firewall, run a cost consolidation exercise. Tally the annual license costs, support costs, and internal engineering hours for each of those three solutions. Then price a full-stack SASE from one of the four leading vendors. The gap is often smaller than expected—and the operational savings are real.
The Accelerating Decline of MPLS
MPLS (Multiprotocol Label Switching) was the dominant enterprise WAN technology for two decades. It offered guaranteed QoS, predictable latency, and carrier-managed reliability—at a price. For enterprises with dozens of locations running mission-critical applications, MPLS was the unquestioned choice for private connectivity.
That consensus has been eroding since SD-WAN began maturing around 2017, and by 2026, it's effectively over for most enterprise use cases. The migration from MPLS to SD-WAN over broadband is no longer a forward-looking initiative—it's a late-cycle cleanup project for most organizations. The vast majority of enterprises that were going to make this transition have made it, or are actively completing it.
Where MPLS is still viable. MPLS still makes sense in a narrow set of scenarios: financial services firms or healthcare organizations with strict regulatory requirements around traffic isolation; manufacturing facilities with real-time control systems that cannot tolerate any packet loss; international enterprise connectivity where broadband alternatives are unreliable or unavailable; and specific last-mile situations where MPLS is actually cheaper than available fiber alternatives (rare, but it happens in some markets).
Where MPLS is not the right choice. For standard enterprise WAN connectivity—offices, retail locations, distribution centers—MPLS is expensive relative to broadband, slow to provision, and inflexible. SD-WAN over dual broadband or broadband plus LTE/5G delivers comparable performance for most workloads at a fraction of the cost, with faster deployment and more granular application control.
If your organization still has large MPLS contracts, the question to ask when those contracts come up for renewal is not "do we want to keep MPLS" but "what specific requirements does MPLS meet that our SD-WAN architecture cannot?" That list should be short—and if it's short, your renewal negotiation should reflect that.
Pull the term dates on your active MPLS circuits. Any circuits renewing in the next 18 months should be flagged for a migration assessment before auto-renewal. Carriers count on passive renewals for MPLS retention—do not give them one without comparing costs against SD-WAN alternatives.
UCaaS Becoming Commoditized
Unified Communications as a Service—the cloud delivery of voice, video, messaging, and collaboration tools—was a meaningful differentiator for early adopters in the early 2010s. By 2026, it is effectively a commodity. Microsoft Teams, Zoom Phone, RingCentral, 8x8, Vonage, Nextiva, and a dozen other providers offer roughly equivalent core functionality at prices that have compressed significantly over the past three years.
Where differentiation still exists. The UCaaS vendors that are not commoditized are the ones competing on specific dimensions that matter to specific buyers: AI features (call intelligence, real-time transcription, agent assist, automated summaries), vertical compliance capabilities (HIPAA-compliant call recording and archiving for healthcare; SEC/FINRA-compliant recording and e-discovery for financial services), and deep CX integrations that turn UCaaS into a customer experience platform. If you're in healthcare, financial services, or running a contact center, these differentiators are real and worth paying for.
If you're a standard enterprise buyer. If you're a multi-location business running standard voice and video calls—not a contact center, not in a regulated vertical—you are in the commodity zone. You should be paying commodity prices. Price pressure is real: many buyers who signed three-year UCaaS contracts in 2021–2022 are renewing into a market where per-seat rates have dropped 20–30%. Use that leverage.
Be careful with long contract terms. Given where UCaaS pricing is headed—down, with AI bundling as a differentiator increasing over the next 18–24 months—locking into a 36-month contract today at current rates may not be wise. Negotiate for 12–24 month terms with renewal options. The vendors will push back; that's normal. But the buyer with pricing leverage in a commoditized market should use it.
If your UCaaS contract renews in the next 12 months, get competitive quotes from at least two other vendors before you engage your current provider on renewal. You don't have to switch—but having a real quote in hand is the only way to negotiate a meaningful price reduction with your incumbent. UCaaS vendors discount heavily to retain accounts; they do not discount simply because you asked.
Carrier Consolidation and What It Means for Buyers
The major carrier market has been consolidating for years. AT&T's acquisition of Warner Media (and its subsequent unwind), T-Mobile's acquisition of Sprint, Lumen's divestitures, Frontier being acquired by Verizon—the list of transactions is long, and the net result is fewer independent carriers competing for enterprise business.
At the mid-market and regional level, the consolidation is even more pronounced. Many of the regional CLECs and competitive carriers that provided meaningful pricing pressure in the 2010s have been acquired, merged, or exited markets. The MSP and telecom agent community has also consolidated, meaning that the broker or agent you bought through five years ago may now be owned by a company with different incentive structures.
Implications for contract leverage. With fewer carriers competing for each market, buyers have less inherent leverage than they did a decade ago. This makes it more important to be methodical: understand your contract terms before they expire, get quotes from all available providers in your market (not just the incumbents), and know what switching actually costs so you can negotiate from an informed position.
Support quality deterioration. Carrier consolidation has not improved support quality. Businesses that previously had dedicated account teams at mid-market carriers often find themselves in generic support queues after acquisitions. Escalation paths that previously existed through personal relationships have been bureaucratized. If you are experiencing support quality problems with your current carrier, document them systematically—this documentation is valuable leverage when you negotiate your next contract or evaluate alternatives.
How to use consolidation as a buyer. The carriers that are acquiring others are often willing to offer aggressive pricing to retain or win large multi-location accounts during integration periods—they need to demonstrate synergies to their leadership. The 12–18 months following a major acquisition is often the best time to negotiate. Use the disruption to your advantage.
Check whether any of your primary carriers have been involved in a merger or acquisition in the past 24 months. If they have, request a review meeting with your account team and come prepared with a competitive benchmark. Integration periods create pricing flexibility that normal contract cycles do not.
What IT Leaders Should Do in 2026
The trends above are not all equally urgent, but they share a common thread: inaction has a cost. Carriers are retiring infrastructure, contracts are auto-renewing at unfavorable rates, and technology consolidation is happening whether or not you participate in it strategically. Here is a prioritized set of actions for IT leaders and their teams:
1. Audit your POTS lines immediately. This is the most time-sensitive action item on this list. Get a complete list of every POTS (analog copper) line in your telecom inventory, document what each line is connected to, and start your replacement planning. If you haven't done this, do it this quarter. The businesses that wait for a carrier discontinuation notice will be making rushed, expensive decisions under a hard deadline.
2. Review every MPLS contract before it auto-renews. Flag every MPLS circuit with a renewal date in the next 24 months. Build a migration assessment for each. Even if you decide to renew MPLS for specific circuits, do it as an intentional choice—not because the contract rolled over passively while you were busy with other things.
3. Assess your SD-WAN and SASE readiness. If you're still running legacy VPN with branch connectivity held together by aging firewalls and point-to-point circuits, you are carrying both security debt and cost debt. A structured assessment of your current architecture against a SASE target state—including a real total cost of ownership comparison—is a reasonable quarter of work for a mid-size IT team or a good scoping project for an outside advisor.
4. Don't over-lock into UCaaS contracts. Given the continued price compression in UCaaS and the rapid development of AI features that will reshape vendor differentiation over the next two years, committing to 36-month contracts at today's per-seat rates is not in your interest. Push for shorter terms, annual pricing reviews, or technology refresh clauses that let you incorporate new AI features without paying for them as an upsell.
5. Benchmark annually, not at contract renewal. The biggest leverage problem in enterprise telecom is that most businesses only look at competitive pricing when a contract is about to expire. By that point, the incumbent knows you're under pressure to make a decision. Build an annual benchmarking process—even a lightweight one—so you always know roughly where market prices are. When your contract comes up, you're negotiating from knowledge, not starting from zero.
6. Evaluate AI in your call routing and UCaaS. Not because AI is a trend worth chasing, but because the ROI calculations on conversational AI for inbound call handling are real and demonstrable. Get the data from your current provider: how many calls are you handling, what's your current containment rate (calls resolved without agent involvement), and what's your average handle time. Those three numbers define your AI opportunity size.
7. Watch the 5G FWA market for your hardest connectivity problems. Don't invest in 5G broadly. Do invest in understanding whether the two or three connectivity challenges you haven't been able to solve with fiber or cable broadband have 5G solutions available now. The FWA market is maturing fast, and for specific sites, it may be the best option you have.
In telecom, passive procurement is expensive procurement. Every trend on this list—AI, POTS elimination, 5G, SASE, MPLS decline, UCaaS commoditization, carrier consolidation—rewards businesses that engage proactively and penalizes those that wait for a carrier to tell them what to do next. The carriers are not your advocates; they're selling you something. The decisions about your architecture and your contracts belong to you.
Want a Second Opinion on Your Telecom Strategy?
ITG Group works with IT leaders across industries to assess their telecom architecture, benchmark contracts against current market rates, and plan transitions for POTS elimination, MPLS migration, and SASE adoption. No carrier relationships to protect. No upsell agenda. Just independent advice.
Start a ConversationFrequently Asked Questions
Is MPLS really dead?
Not completely, but it's dead for most standard enterprise use cases. MPLS still makes sense in a narrow set of scenarios: regulated industries with strict traffic isolation requirements, real-time manufacturing control systems, and some international connectivity situations. For typical enterprise WAN connectivity—offices, retail locations, distribution centers—SD-WAN over broadband outperforms MPLS on cost and flexibility without meaningful sacrifices in reliability for most workloads. If you're asking whether to renew an expiring MPLS contract, the answer is almost certainly to run a migration assessment first. The default answer should no longer be "yes, renew."
What do I do about my POTS lines?
Start with an inventory. Pull every POTS line from your telecom invoice and document what each one is connected to—a phone system, an alarm panel, a fax machine, an elevator, a building controller. Once you know what each line does, you can match it to the right replacement: SIP trunking for voice lines, cellular backup devices or POTS-to-IP gateways for alarm and security systems, cloud fax for dedicated fax lines, and modern SIP-compatible hardware for building controls and elevator phones. Do not accept your carrier's recommended replacement without evaluating alternatives—their solution may be right for some lines and wrong for others. Start now; don't wait for a discontinuation notice.
Is AI in UCaaS worth paying for yet?
It depends entirely on your use case. For contact centers and call-heavy businesses, AI features—conversational IVR, real-time agent assist, automated call summaries, AI-driven quality monitoring—deliver measurable ROI and are worth paying for. For standard office voice and video collaboration, the AI features being bundled into platforms like Microsoft Teams and Zoom are often already included in your existing licensing. The question is whether you're configured to use them, not whether you should pay extra for them. Where AI is clearly not worth paying extra for yet: niche AI features sold as premium add-ons by smaller UCaaS vendors with unproven models and no track record. Ask for data on call containment rates and handle time improvements before you commit to paying for AI features.
What is SASE, and do we need it?
SASE (Secure Access Service Edge) combines SD-WAN connectivity with cloud-delivered network security—zero-trust network access, secure web gateway, cloud access security broker, and next-gen firewall—into a single vendor platform. Whether you need it depends on your current architecture. If you're running a separate SD-WAN platform, a separate remote access VPN, and separate next-gen firewall licensing, you are already spending money on the components of SASE. The question is whether consolidating them under one full-stack SASE vendor would reduce complexity and cost. For many mid-market and enterprise organizations, the answer is yes—especially given how remote work has changed the perimeter. A structured SASE assessment is a reasonable place to start.
How do I negotiate with carriers when there are fewer of them?
Fewer carriers does not mean no leverage—it means you have to be more methodical about creating leverage. The most effective tactics: get competitive quotes from every available provider in your market, not just the incumbents; know your switching costs before you negotiate; document support quality failures as contract negotiation evidence; and time your negotiations so that you are engaging well before contract expiration, when the carrier has real motivation to retain you. Post-acquisition periods—when a carrier has just acquired a competitor—are often unusually good times to negotiate, because acquired carriers are under pressure to demonstrate retention metrics. Use that window.
How does ITG Group help with telecom strategy?
ITG Group works as an independent telecom advisor—not a carrier, not an agent with exclusive relationships. We help businesses inventory their telecom environment, benchmark contracts against current market rates, evaluate technology transitions (MPLS to SD-WAN, POTS replacement, UCaaS migration, SASE architecture), and negotiate with carriers from a position of knowledge rather than pressure. We're paid for our advisory work, not by carrier commissions, which means our recommendations reflect your interests. If you're navigating any of the trends on this page—POTS retirement, MPLS renewal, SASE evaluation—a conversation with ITG is a good first step.