Quick answer
Copper carried its own power from the exchange, so a corded phone worked in a blackout. Nothing that replaces it does. Digital voice needs mains power and a working internet connection, so resilience has to be designed rather than assumed - and the good news is that designing it is cheap. A small uninterruptible power supply on the router, an automatic divert of your main number to mobiles, and mobile data as a backup connection cover almost every realistic failure. Do those three and you end up more reachable than you were on copper, because a cloud-hosted number can be answered anywhere while a cut cable could not.
The reason this matters more than the copper nostalgia suggests: your phones, your card payments and possibly your alarm signalling now share one connection. That concentration is new, and it deserves a plan. If you want one designed for your building, get a no-obligation quote.
What copper actually gave you, honestly
It is worth being accurate about the thing we are replacing, because the comparison usually gets overstated in both directions.
Copper delivered a small current from the exchange, so a basic corded phone kept working when your building lost power. That was genuinely valuable. But it did not cover as much as people remember: cordless phones never worked in a power cut, on-site phone systems needed mains power, and anything with a display or an answering machine was dead too. Plenty of businesses discovered during an outage that the only working phone was one nobody had used in years.
Copper also had a failure mode nothing could work around. Cut the cable, flood the exchange or damage the cabinet, and you were off - and there was no rerouting a physical pair. So the fair comparison is not "copper worked, digital does not". It is one line with one narrow advantage against a system that can be designed to reroute around almost anything, if you design it.
The four failure modes to design for
| Failure | How likely | What breaks | The layer that covers it |
|---|---|---|---|
| Short power cut | Common | Router, ONT, handsets, so all calls | UPS on the network kit |
| Extended power cut | Occasional | Everything, including the UPS eventually | Automatic divert to mobiles |
| Broadband fault | Occasional | Calls, but power is fine | 4G or 5G failover, or divert |
| Equipment or router failure | Occasional | Everything at that site | Divert, plus a spare router |
| Building inaccessible | Rare, but total | Everything on site | Cloud routing to home or mobiles |
Most resilience plans stop after row one, because a UPS is a thing you can buy and the rest requires thought. Rows two to five are where the real protection is, and they are mostly configuration rather than hardware.
Layer one: battery backup, sized properly
A small uninterruptible power supply keeps your router, the ONT and your network switch running through a short outage. It is the cheapest meaningful improvement available and it takes an hour to fit.
Two things to get right. Put everything in the chain on it. A UPS on the router alone is useless if the ONT or the network switch that powers your handsets is on an unprotected socket - and this is the mistake we see most. Trace the whole path from the fibre entry to the desk and protect every powered device on it.
Be realistic about runtime. A UPS sized for the network kit alone can hold up for a decent stretch; add a rack of servers and it will not. Work out what you actually need to keep alive - the connection and the phones - and leave the rest to shut down gracefully. If your handsets are powered over Ethernet from a switch, that switch is part of the phone system for this purpose.
Layer two: the divert rule that fires by itself
This is the layer that does the most work and costs the least, and it is routinely left out because it is invisible.
On a hosted service, your number lives in the provider's platform rather than in your building. That means the platform can notice that your handsets have stopped responding and redirect calls somewhere else - to a group of mobiles, to another site, to voicemail-to-email - without anybody at your end doing anything.
The distinction that matters is between a divert you have to switch on and a rule that fires automatically. In a real outage the person who knows how to switch it on is stuck in traffic, or the office is dark and nobody can find the laptop. Configure the automatic version, then decide three things:
- Who the calls go to, and in what order. A group of two or three mobiles beats one.
- What callers hear if nobody answers - a message acknowledging the outage is better than a generic voicemail.
- How it turns off again, so calls come back to the office when things recover.
This is the heart of business continuity with hosted telephony.
Layer three: a second path to the internet
If your broadband fails but the power is fine, a backup connection keeps everything running rather than just the phones. In practice this means a router with a mobile data failover - a 4G or 5G backup that takes over automatically when the primary drops.
Whether it is worth it comes down to what else shares the connection. When it was only email, an hour offline was an inconvenience. Now the same connection may carry your phones, your card payments and possibly your alarm signalling path, and an hour offline is three services down at once. That concentration is the honest argument for failover, and it is a much stronger one than it used to be.
Emergency calls in an outage
Say this out loud to your team rather than leaving it to be discovered.
999 and 112 do not work from a digital phone during a power cut, because nothing in the chain is powered. Ofcom's protections in this area are framed around residential and vulnerable customers rather than businesses, so do not assume a battery back-up unit will be offered with a business service - ask your provider what they provide, and get the answer in writing.
For most businesses the practical answer is straightforward and should be explicit in your procedures: a charged mobile is the emergency route when the power is off. Make sure at least one person on site always has one, and that everybody knows that is the plan.
One related detail: the address registered against your voice service is what gets passed to the emergency operator. Check it is correct when the service goes live, and update it if you move.
Safety-critical devices are a separate problem
Resilience planning for handsets is not resilience planning for the building.
A lift emergency phone must work while somebody is trapped in a lift during a power cut, which is precisely the scenario your office broadband cannot survive. That is why the standard answer is a self-contained mobile unit with its own battery, fitted by your lift maintenance company - not an adapter on your network.
The same logic applies to monitored alarm signalling, telecare and anything else where failure has safety or insurance consequences. Each needs its own path, arranged with its own supplier. Start from devices that rely on phone lines and do not let a phone migration quietly absorb them.
The test nobody runs
A resilience plan that has never been exercised is a document, not a capability. Book an hour, pick a quiet morning, and actually do this:
- Pull the power to the router and network kit. Time how long calls keep working.
- Leave it off past the UPS runtime. Confirm the automatic divert fires, and to whom.
- Ring the main number from an outside mobile and check what the caller hears.
- Check every published number diverts, not just the main one.
- Restore power and confirm calls return to the office without manual intervention.
- Separately, unplug the broadband and confirm the mobile failover takes over.
- Write down what happened, what surprised you, and who fixed it.
The two findings that come out of this test most often: something important was on an unprotected socket, and the divert went to somebody who left the company. Both are five-minute fixes once you know.
Multi-site: the single point of failure you did not design
One thing worth flagging for businesses with more than one location. Consolidating onto one platform is usually right, but check whether your sites now depend on a single connection or a single router somewhere. Sites can back each other up - calls for a site that is down can route to another one automatically, which is a genuinely strong position - but only if the design allows for it. See multi-site PSTN migration.
The bottom line
Yes, digital phones need power and internet, and that is a real change worth planning around. It is also a small, cheap problem with well-understood answers: protect the whole powered chain with a UPS, configure a divert that fires by itself, add mobile data failover if the connection carries more than email, keep a charged mobile for emergencies, and give safety-critical devices their own path. Do that and the honest comparison is not close - one copper pair could never reroute around anything.
Want your resilience designed and then actually tested rather than assumed? Get a no-obligation quote or call 0333 006 9399.
Getting unknown calls on a business line? Our free UK phone number checker shows which network or provider Ofcom allocated any number to, and whether the range is in service — no sign-up.
Frequently asked questions
Will my phone work in a power cut after the switch-off?
Not by default. Digital phone lines need mains power for the router, the ONT and the handsets, so they stop when the power does. A UPS on the network kit covers short outages, and an automatic divert of your number to mobiles covers longer ones. Both need setting up before you need them.
Can I call 999 from a digital phone line in a power cut?
No. Nothing in the chain is powered, so emergency calls stop along with everything else. Ofcom's protections here are framed around residential and vulnerable customers rather than businesses, so ask your provider what they offer. Assume a charged mobile is your emergency route and tell staff that explicitly.
What size UPS do I need for my phone system?
Size it for the whole powered chain rather than one box - the ONT, the router, the network switch that powers your handsets, and the handsets themselves if they are not powered over Ethernet. The common mistake is protecting the router while leaving the ONT or switch on an unprotected socket.
Is a digital phone line less reliable than copper?
Only in the narrow sense that copper powered a corded phone during a blackout. In every other respect a cloud-hosted number is more resilient, because it can reroute to mobiles or another site automatically, whereas a cut copper pair could not be rerouted at all. The difference is that digital resilience has to be configured.
What is the most important thing to set up for phone resilience?
The automatic divert rule. Hardware protects you for as long as the batteries last; a rule that redirects calls to mobiles without anybody triggering it protects you indefinitely and costs nothing. Configure it so it fires on its own, because in a real outage nobody will be at a laptop.
Do I need 4G or 5G failover for my phones?
It depends on what else shares the connection. If phones, card payments and alarm signalling all ride on one broadband service, a single fault takes out three things and failover is easy to justify. If the connection only carries phones and a divert to mobiles is acceptable, you may not need it.
What about my alarm or lift line in a power cut?
Those need their own resilient path and should never sit behind office broadband. A lift emergency phone in particular must work while somebody is trapped during a power cut, which is why the standard answer is a self-contained mobile unit with integral battery, fitted by your lift maintenance company.
