How Do You Move an Office to VoIP Without Downtime? (w/Examples) + FAQs

You move an office to VoIP without downtime by running your old phone system and the new VoIP system in parallel, porting numbers in small batches, and cutting over during off-hours after a full bandwidth, E911, and user-training check. A zero-downtime cutover depends on careful number porting under the FCC’s local number portability rules, proper E911 setup under RAY BAUM’s Act, and a bandwidth plan that supports every concurrent call your team makes.

The real problem is that phone service is mission-critical, and even one hour of dead lines can cost sales, patient bookings, legal deadlines, or customer trust. Federal rules like Kari’s Law, the RAY BAUM’s Act Section 506, and the FCC’s porting timeline under 47 CFR § 52.36 set the rules for how you cut over, and breaking them creates fines, failed 911 calls, and lost numbers.

According to a 2024 Metrigy survey on UCaaS migration, 73% of businesses that switched to VoIP reported at least some downtime during cutover, and the average outage ran 4.2 hours. You can beat that average with the right plan.

Here’s what you’ll learn in this guide:

• 📞 How to port numbers in phases so calls never drop
• 🌐 How to size bandwidth and QoS so voice stays clear under load
• 🚨 How to meet Kari’s Law and RAY BAUM’s Act E911 rules on day one
• 🧪 How to run a parallel pilot that catches problems before cutover
• 🛠️ How to avoid the seven most common mistakes that cause outages

Why Moving to VoIP Is Risky Without a Plan

VoIP (Voice over Internet Protocol) sends calls as data packets over your internet connection instead of copper wires. The technology itself is mature, but the switch from a legacy PBX or landline to VoIP touches carriers, numbers, routers, firewalls, handsets, 911 records, and user habits all at once. When any one piece lags, the phones go dark.

Federal law treats phone service as a protected utility in many ways. The FCC’s Part 52 rules force carriers to port numbers on request, but the process has a standard window of one business day for simple ports and up to 20 business days for complex ones. If you miss a porting step, your number can stall in “limbo” where neither carrier owns it, and inbound calls fail.

The consequence of a sloppy cutover goes beyond lost calls. Under Kari’s Law, every multi-line phone system must allow direct 911 dialing without a prefix, and failure to comply can bring FCC fines up to $10,000 plus $500 per day. A common misconception is that Kari’s Law only applies to hotels; in fact, it applies to every multi-line telephone system installed, manufactured, or sold after February 16, 2020.

The Core Components of a VoIP Cutover

A clean cutover has six moving parts, and each one has its own failure mode. The first is the LOA (Letter of Authorization), which tells your new carrier it has permission to pull your numbers. The second is the CSR (Customer Service Record), which must match your billing name and address exactly or the port rejects.

The third part is bandwidth and QoS (Quality of Service) on your LAN. The fourth is E911 address registration for every extension. The fifth is handset or softphone provisioning. The sixth is user training, which is the piece most teams skip and then regret.

The consequence of ignoring any one of these is a partial outage. For example, if you port numbers before QoS is tuned, calls connect but sound garbled, and customers hang up. A common misconception is that “the carrier handles everything” — in reality, the carrier handles the number, but you handle the network and the endpoints.

The Zero-Downtime Playbook: Parallel Running

The single most important rule is to keep your old system live until the new system proves itself. This is called parallel running, and it is how large enterprises and careful small businesses avoid outages. You forward your main number to a temporary VoIP DID (Direct Inward Dial) during testing, then port the real number only after the pilot group signs off.

Under 47 CFR § 52.36, a simple port should take one business day once the carrier accepts the LOA. The consequence of skipping parallel running is that any port rejection, firewall bug, or DNS issue becomes a live outage instead of a test finding. A real-world example: Alvarez & Chen, a 12-lawyer firm in Denver, kept their Comcast lines active for 21 days while testing RingCentral, then cut over on a Friday at 6 p.m. after the last court filing.

A common misconception is that parallel running doubles your phone bill forever. In practice, you pay for both services for two to six weeks, which is cheap insurance compared to the revenue lost in a half-day outage.

Step-by-Step Cutover Sequence

The sequence matters more than the speed. Step one is to audit every number, extension, hunt group, and auto-attendant on the old system. Step two is to order the new VoIP service and request temporary DIDs so users can test without touching the live numbers.

Step three is to tune the network: enable QoS on the router, open SIP (Session Initiation Protocol) ports per the provider’s spec, and run a VoIP readiness test from tools like the 8×8 Network Readiness Test or a speed and jitter test. Step four is to register every E911 location in the carrier portal, which satisfies Section 506 of RAY BAUM’s Act. Step five is a pilot with five to ten power users for one to two weeks.

Step six is the port itself, scheduled for a low-traffic window such as Friday evening or a holiday. Step seven is a 48-hour hypercare period where IT watches call logs, MOS (Mean Opinion Score) scores, and help-desk tickets. A common misconception is that you can do all seven steps in a weekend; for anything larger than ten seats, give yourself four to eight weeks.

Bandwidth, QoS, and Network Readiness

Voice traffic is small but unforgiving. A single G.711 call uses about 87 kbps in each direction, and a G.729 call uses about 32 kbps. The consequence of under-provisioning is packet loss, jitter, and dropped calls, which customers notice within seconds.

You size bandwidth by multiplying your busy-hour concurrent calls by the codec rate and adding 20% overhead. For a 30-person office where 40% of staff are on calls at peak, that is 12 concurrent calls × 100 kbps × 1.2 = about 1.5 Mbps each way dedicated to voice. A real-world example: Dr. Patel’s dental group in Mesa, Arizona ran fine on 50/10 Mbps cable until they added a second location; at that point jitter spiked, and they upgraded to a 100/100 Mbps fiber circuit from Cox Business.

A common misconception is that “fast internet” alone is enough. In reality, VoIP needs low jitter (under 30 ms), low packet loss (under 1%), and low latency (under 150 ms round-trip), which require QoS rules on the router and a clean LAN, not just raw speed.

Configuring QoS and Firewalls

QoS tags voice packets with DSCP (Differentiated Services Code Point) value 46 / EF, which routers and switches then push to the front of the queue. Without QoS, a large file upload can starve voice packets and cause a dropped call. The consequence is angry customers and confused staff who blame the VoIP vendor for a LAN problem.

Firewalls add another layer. Most VoIP providers need UDP ports 5060 and 5061 for SIP, plus a range like 10000–20000 for RTP (Real-time Transport Protocol) media. You must also disable SIP ALG (Application Layer Gateway) on consumer-grade routers, because SIP ALG rewrites packet headers in ways that break modern VoIP. A common misconception is that “more security is always better” — for VoIP, aggressive SIP ALG is the number-one cause of one-way audio.

Number Porting Without Losing a Call

Porting is the transfer of a phone number from one carrier to another. Under the FCC’s local number portability rules, carriers must accept valid port requests and cannot hold numbers hostage. The process is standardized but still fragile if you make a mistake on the paperwork.

A clean port needs four items that match exactly: the account number on the losing carrier, the billing telephone number (BTN), the authorized user’s name, and the service address. The consequence of a mismatch is a port rejection, which can delay cutover by 7 to 10 business days. A real-world example: Greenleaf Nonprofit in Austin, Texas listed its service address as “Suite 200” on the LOA, but AT&T had “Ste. 200” on the CSR; the port rejected twice before the ED fixed the abbreviation.

A common misconception is that you should cancel your old service before porting. Never do this. Canceling releases the number back to the carrier pool, and you lose it forever.

Scenario Tables: How Cutovers Go Right or Wrong

E911 Compliance: Kari’s Law and RAY BAUM’s Act

Every VoIP cutover must meet two federal 911 rules. Kari’s Law, named after Kari Hunt who was killed in 2013 when her daughter could not reach 911 from a hotel phone, requires every multi-line telephone system to allow direct 911 dialing with no “9” prefix. The law also requires a simultaneous on-site notification to a front desk, security, or manager when 911 is dialed.

RAY BAUM’s Act Section 506 adds the “dispatchable location” rule. Each 911 call must send a street address, building, floor, and room or suite number accurate enough for first responders to find the caller. The consequence of non-compliance is FCC fines and, far worse, a delayed emergency response.

A real-world example: Thompson Logistics in Columbus, Ohio rolled out Microsoft Teams Phone to 85 employees but forgot to set per-user emergency addresses for remote workers; when a remote employee dialed 911, dispatch was sent to the Columbus HQ instead of the worker’s home in Dayton. A common misconception is that cloud VoIP handles 911 automatically. In reality, you must register each physical location and, for softphone or mobile users, prompt them to confirm their location at login.

Mapping Every Extension to a Location

For a fixed-desk office, you register the building address once and assign every DID to that location. For a multi-floor office, you break it into emergency response locations (ERLs) by floor or wing so responders find the caller fast. For remote or hybrid workers, the softphone app must prompt for a current address and update it when the user moves.

The consequence of skipping ERL mapping in a large building is that paramedics land in the lobby while the patient is on the fifth floor. A common misconception is that ERLs are only for big buildings; the FCC guidance says any MLTS where a generic address would delay response must use dispatchable location data.

Security, Encryption, and Compliance

VoIP traffic crosses the public internet, so it must be encrypted in transit. The industry standard is SIP over TLS (Transport Layer Security) for signaling and SRTP (Secure Real-time Transport Protocol) for media. Without encryption, a bad actor on the same network can record calls or harvest credentials.

NIST Special Publication 800-58 lays out the baseline for VoIP security in U.S. organizations, including network segmentation, strong authentication, and regular patching. The consequence of ignoring NIST guidance is a failed audit for HIPAA-regulated clinics, PCI-DSS retailers, or CJIS-bound law-enforcement vendors. A real-world example: a 40-bed clinic in Tampa chose a low-cost VoIP reseller that did not offer SRTP; their next HIPAA risk assessment flagged the unencrypted calls as a reportable gap.

A common misconception is that VoIP is inherently less secure than landlines. In practice, a properly configured VoIP system with TLS, SRTP, and STIR/SHAKEN caller-ID authentication is more secure than a copper line that anyone with alligator clips can tap.

STIR/SHAKEN and Robocall Rules

Under the TRACED Act and FCC STIR/SHAKEN rules, originating carriers must sign outbound calls with a cryptographic token that proves the caller-ID is legitimate. If your new VoIP provider does not support STIR/SHAKEN at attestation level A, your outbound calls may be flagged as “Spam Likely” on recipient phones. The consequence is a sharp drop in answer rates, which kills sales teams.

A common misconception is that STIR/SHAKEN only matters for call centers. Any business that places outbound calls benefits from A-level attestation, so ask every VoIP vendor to confirm their attestation level before signing.

Real-World Examples of Zero-Downtime Cutovers

Three named examples show how the playbook works in different industries. Each example highlights a different pain point and the fix that preserved uptime.

Example 1: Maria’s Law Firm Relocation

Maria Alvarez runs a 12-attorney firm in Denver that was moving from a leased suite on 17th Street to a new building on Larimer. Her goal was zero dropped calls during a two-week move window. She ordered Nextiva VoIP service six weeks ahead, provisioned temporary DIDs, and trained every attorney on the desktop and mobile apps while the old NEC PBX still ran in the 17th Street office.

On moving weekend, Maria kept the old PBX powered on at the old suite (landlord cooperation) and forwarded the main line to a Nextiva DID at 5 p.m. Friday. The port FOC (Firm Order Commitment) hit Monday at 9 a.m., and no inbound call was lost. The consequence of her parallel-running plan was a Monday morning that felt normal, even though every piece of infrastructure had changed.

Example 2: Dr. Patel’s Multi-Site Clinic

Dr. Raj Patel operates three dental offices across Mesa, Chandler, and Gilbert, Arizona. His problem was that patients call a single 480 number that routes to whichever office has the next available chair. He picked RingCentral MVP because its cloud PBX handles multi-site hunt groups without an on-prem appliance.

Dr. Patel ported the main 480 number last, after porting each location’s direct line first. The consequence of this sequence was that any porting issue would hit a single location, not the master routing number. He also tagged every handset with an E911 location ID per RAY BAUM’s Act, so a 911 call from Gilbert reaches the Gilbert PSAP, not Mesa.

Example 3: Greenleaf Nonprofit’s Remote-Hybrid Cutover

Greenleaf Nonprofit in Austin, Texas has 28 staff, half of whom work from home. Their executive director, David Okafor, chose Zoom Phone because the team already used Zoom Meetings. The problem was that remote workers’ home Wi-Fi varied widely in quality.

David ran a two-week pilot with the development team using Zoom’s built-in network quality reports. The consequence of the pilot was three clear findings: two staff needed wired connections, one needed a better headset, and one had a SIP ALG router that had to be replaced. Cutover weekend passed with zero tickets because the problems were fixed before go-live.

Mistakes to Avoid

Even with a good plan, small errors cause big outages. Here are the seven most common mistakes and the fix for each.

• Canceling old service before porting completes. The number drops back to the carrier pool and is gone forever; always let the port FOC close first.
• Skipping the Customer Service Record check. A name or address mismatch kills the port; pull the CSR from the losing carrier and match it byte-for-byte.
• Porting everything in one batch. A single rejection delays hundreds of numbers; port in waves of 10 to 25.
• Forgetting to disable SIP ALG on the router. SIP ALG rewrites headers and causes one-way audio; turn it off on every router and firewall.
• Under-sizing bandwidth. Busy-hour concurrent calls × codec rate × 1.2 is the minimum; anything less causes jitter and drops.
• Ignoring E911 registration. Dispatchable location is required under RAY BAUM’s Act; register every site and every remote worker.
• Skipping user training. Staff who don’t know how to transfer, park, or page will blame the new system; run 30-minute sessions before cutover.

Do’s and Don’ts of a VoIP Office Move

The following do’s and don’ts capture the habits that separate clean cutovers from painful ones.

Do’s:

• Do run parallel systems for at least two weeks, because pilots catch problems before they become outages.
• Do tune QoS to DSCP 46 for voice, because voice packets must beat file uploads to the router queue.
• Do register every E911 location, because Kari’s Law and RAY BAUM’s Act require it on day one.
• Do use SIP over TLS and SRTP, because unencrypted VoIP fails HIPAA and PCI audits.
• Do keep the old carrier’s contact info handy, because a fast un-port back is your rollback plan if things go wrong.

Don’ts:

• Don’t cancel the old service before FOC, because a canceled number cannot be recovered.
• Don’t use consumer-grade Wi-Fi for handsets, because roaming and interference cause drops that look like VoIP failures.
• Don’t cut over on a Monday morning, because any issue hits peak traffic and multiplies the damage.
• Don’t rely on carrier sales promises, because contract SLAs are what actually bind the provider.
• Don’t forget about fax, alarm, and elevator lines, because these often need POTS-replacement devices and special ATAs rather than standard SIP.

Pros and Cons of Moving an Office to VoIP

VoIP is a clear win for most offices, but it is not free of trade-offs. Here is the honest balance sheet.

Pros:

• Lower monthly cost, because VoIP plans bundle unlimited U.S. calling at $15–$35 per seat versus $50+ per PRI channel.
• Work-from-anywhere flexibility, because softphones and mobile apps turn any broadband connection into an office extension.
• Built-in features like call recording, analytics, and SMS, because cloud PBXes include what legacy systems charged extra for.
• Easier scaling, because adding a seat is a portal click instead of a new copper pair.
• Better integrations with CRM and help desk tools, because modern VoIP APIs plug into Salesforce, HubSpot, and Zendesk.

Cons:

• Internet dependency, because a broadband outage takes every line down unless you have LTE failover.
• Power dependency, because handsets need PoE or local power, so UPS backup matters.
• Quality depends on LAN tuning, because bad QoS causes jitter that copper never had.
• 911 complexity, because dispatchable location for remote users requires ongoing attention.
• Feature parity gaps, because some legacy PBX features like overhead paging or complex call-center routing may need extra hardware or licensing.

Processes, Forms, and Carrier Paperwork

Every VoIP cutover touches a handful of standard forms. The LOA (Letter of Authorization) is the master document; it must list every number to port, the losing carrier’s name, the account number, the BTN, and the authorized signer. A single wrong digit rejects the port.

The CSR (Customer Service Record) is pulled from the losing carrier and shows exactly what is on file. You match the LOA to the CSR, not to your own assumptions. The FOC (Firm Order Commitment) is the date and time the port will complete; under FCC rules, simple ports get a one-business-day FOC, and complex ports get up to 20 business days.

The E911 registration form lives inside your new VoIP provider’s portal, and it captures the dispatchable location for every DID. The consequence of skipping a single DID is that a 911 call from that extension may route to a default address. A common misconception is that you can submit the LOA and E911 forms after cutover; in reality, both must be complete before the port executes.

Step-By-Step Paperwork Flow

Step one: collect the losing carrier’s most recent bill to capture the exact account name, number, and address. Step two: have the new provider pull the CSR. Step three: resolve any mismatches before signing the LOA, because every mismatch is a guaranteed rejection.

Step four: sign the LOA and submit with the port order. Step five: receive the FOC date and communicate it to all stakeholders. Step six: complete E911 registration for every DID and confirm test calls to 933 (a non-emergency test number many carriers provide).

The consequence of skipping step three is a two-week delay per rejection. A common misconception is that you can “fix it later” if the port fails; each failure resets the clock.

Key Entities in a VoIP Migration

Several organizations set the rules or provide the services that shape a VoIP cutover. The Federal Communications Commission (FCC) sets porting, 911, and robocall rules. The North American Numbering Plan Administrator (NANPA) manages area codes and number pools.

The iconectiv NPAC (Number Portability Administration Center) is the clearinghouse that tracks which carrier owns each ported number. The Alliance for Telecommunications Industry Solutions (ATIS) publishes the standards behind STIR/SHAKEN and other VoIP protocols. Your losing carrier is the company you are leaving, and your gaining carrier (or CPaaS/UCaaS vendor) is the new provider.

Each entity has a role, and they relate in a chain: the FCC sets rules, NANPA and iconectiv run the number infrastructure, ATIS defines the standards, and the carriers execute against those standards. The consequence of not knowing who does what is wasted time — for example, calling the FCC about a port dispute when the real resolution path is carrier-to-carrier escalation through the NPAC.

Relevant FCC Rulings and Precedents

Several FCC orders shape today’s VoIP landscape. The 2007 Local Number Portability Order cut the standard porting interval to one business day for simple wireline and wireless ports. This order is the reason you can now expect fast ports instead of the old 5-day default.

The 2019 Kari’s Law implementation order and the 2020 RAY BAUM’s Act Section 506 order together created the direct-dial-911 and dispatchable-location rules for MLTS systems installed after February 16, 2020. The consequence of non-compliance is civil penalties plus the human cost of delayed emergency response.

The TRACED Act of 2019 and the FCC’s STIR/SHAKEN implementation order require caller-ID authentication for IP-based voice providers. A common misconception is that these rules only bind big carriers; in fact, any VoIP-originating business should verify its vendor signs calls at attestation level A.

State Nuances to Watch

State public utility commissions add wrinkles on top of federal rules. California’s CPUC regulates VoIP interconnection fees and 911 surcharges. Texas’s PUC requires certain business lines to maintain copper-based backup for alarm monitoring in specific industries.

New York’s PSC has a VoIP consumer-protection framework that affects how providers disclose outages. Illinois requires ICC registration for certain CLECs and interconnected VoIP providers. The consequence of ignoring state rules is unexpected fees and, in some cases, cease-and-desist letters from the state regulator.

A common misconception is that “VoIP is federally preempted” across the board. In reality, the FCC has preempted entry regulation for interconnected VoIP, but states retain authority over consumer protection, 911 fees, and certain taxes, so always check the state PUC site before cutover.

Frequently Asked Questions

Can I keep my existing phone number when I switch to VoIP?

Yes. Federal local number portability rules under 47 CFR § 52 require carriers to port your number to any qualified provider, as long as you remain in the same rate center or the new carrier supports the original rate center.

Will my phones really stay up during the switch?

Yes. If you run the old system and the new VoIP system in parallel, port numbers in waves, and cut over during off-hours, calls stay connected; the FCC’s one-business-day porting standard makes the final transition nearly instant.

Do I need fiber internet for VoIP to work?

No. Any reliable broadband with enough bandwidth and low jitter works; a 30-person office can run well on a 100/20 Mbps cable circuit if QoS is tuned and SIP ALG is disabled.

Is VoIP safe for HIPAA or PCI-regulated businesses?

Yes. VoIP is compliant when the vendor supports SIP over TLS, SRTP media encryption, and a signed BAA for HIPAA, in line with NIST SP 800-58 security guidance for VoIP systems.

Must my VoIP system allow direct 911 dialing?

Yes. Kari’s Law requires every multi-line telephone system installed after February 16, 2020, to allow direct 911 calls without a “9” prefix and to send on-site notification of the call.

Do remote and hybrid workers need special 911 setup?

Yes. RAY BAUM’s Act Section 506 requires a dispatchable location for every 911 call, so softphones must prompt remote users to confirm their current address at sign-in or relocation.

Can porting fail and leave me without a number?

No. A failed port simply stays with the losing carrier; you do not lose the number, you just must fix the LOA or CSR mismatch and resubmit, which is why you should never cancel old service before FOC.

Will VoIP calls sound as good as landlines?

Yes. With QoS set to DSCP 46, jitter under 30 ms, and packet loss under 1%, G.711 or Opus codec calls match or exceed landline clarity per ITU-T MOS benchmarks.

Do I need a separate plan for fax, alarms, and elevators?

Yes. These analog devices often fail on standard SIP trunks and may require POTS-replacement devices, T.38 fax gateways, or dedicated cellular units to stay compliant with elevator codes and alarm UL listings.

Is STIR/SHAKEN something I need to worry about?

Yes. Without A-level attestation under the FCC’s STIR/SHAKEN rules, your outbound calls may be labeled “Spam Likely,” so confirm the provider’s attestation level before signing a contract.

Can I roll back to my old system if VoIP fails after cutover?

Yes. If you kept the old carrier’s service active through parallel running, you can request an emergency un-port back within the first 30 days, which is why early cancellation of the legacy service is the single most dangerous cutover mistake.

How long does a full VoIP migration usually take?

Yes, timelines are predictable: small offices under 25 seats take two to four weeks, mid-market offices take four to eight weeks, and enterprise cutovers of 250+ seats take three to six months with staged porting and phased ERL rollouts.