Every page you will find on this subject is published by somebody selling one. That shapes the material more than most readers notice: features get listed, trade-offs get thinner treatment, and the question of whether you need the thing at all rarely comes up.
TabaTalk does not sell cloud phone systems, which makes this easier to write honestly. What follows covers how the technology works, what it genuinely replaces, where the costs actually sit, and the regional conditions deciding whether a given option is obtainable at all.
What is a cloud phone system?
A cloud phone system is business telephony delivered as a managed subscription, with the call handling running on a provider’s infrastructure instead of equipment installed at your premises. Staff get extensions, voicemail, transfers and hold music through software and apps rather than through a box in a cupboard.
The older arrangement it replaces is the private branch exchange, universally shortened to PBX. That was a physical switch, bought outright, maintained under contract and replaced every seven to ten years. Moving it to the cloud changes the ownership model and the upgrade cycle far more than it changes what a caller experiences.
| Traditional PBX | Cloud phone system | |
| Where it runs | Hardware on your premises | Provider infrastructure |
| How you pay | Capital purchase plus maintenance | Monthly subscription per user |
| Adding people | Engineer visit, possible upgrade | Configuration change |
| Upgrades | Projects every few years | Continuous, handled by the provider |
| Works from home | Rarely, or through add-ons | By default |
| Fails when | Hardware breaks | Connectivity drops |
Vocabulary varies more than the product does. Cloud telephony, hosted PBX, cloud PBX and business VoIP describe broadly the same thing from different angles, and the analyst category covering them is unified communications as a service, shortened to UCaaS. Gartner defines that market as covering telephony, messaging and meetings, with organisations communicating both internally and externally through it.
That last point is worth holding on to. Plenty of published comparisons insist this is purely an internal tool, while the analyst wording runs broader than that.
Why does the terminology sprawl so much? Partly because vendors coming from different starting positions named the same destination differently. Equipment manufacturers called it hosted PBX because they were describing their old box moved elsewhere. Telecoms operators called it cloud telephony because they think in terms of service delivery. Software companies called it UCaaS because subscription categories are how their market is segmented. All three descriptions are defensible, and a buyer reading three proposals can reasonably conclude they cover three products.
They mostly do not. Compare what each actually delivers rather than what anyone calls it, since the labels tell you more about the seller’s history than about the software.
How a cloud phone system works
Calls travel over a broadband link to your provider’s platform, which decides where they should go and delivers them to whichever device is meant to ring. Your caller hears an ordinary phone call throughout, with no indication that anything unusual happened.
The path a call takes
Five stages, worth knowing because each one fails differently:
- Someone dials your number. The call enters the public telephone network as normal.
- It reaches your provider. Whoever holds that numbering range accepts the call and matches it to your account.
- Audio becomes data. Voice is packetised for an IP path in place of a reserved circuit.
- Your rules decide the destination. Business hours, extension dialled, caller identity and availability all feed that decision.
- A device rings. A softphone, desk handset, mobile app or browser tab alerts somebody, who answers.
Stages one through three belong to the provider and the operators upstream. Stage four belongs to you, and there sits nearly all the configurable value.
Outbound runs the same path in reverse, with one addition worth noting. The number presented as your caller identity is a configuration choice, not a fixed property of the line, which is useful for presenting a single company number from any device and consequential enough to attract regulation in several markets.
What lives where now
Under the old arrangement, almost everything was on site: the switch, the line cards, the voicemail store, the configuration. A technician visited to change anything meaningful.
Cloud delivery moves the intelligence off site and leaves only endpoints behind. Handsets, headsets, connectivity and possibly a network switch, and that is the extent of it. The configuration lives in a browser, which is why enrolling somebody new takes minutes instead of a scheduled visit.
Two consequences follow that matter commercially. Your dependency moves from hardware reliability to connectivity reliability, which changes what you should be spending on resilience. And the skills you need in-house change too, from telephony engineering toward network management and administration.
Neither is worse than what came before. Both are different enough that plans made on the old assumptions tend to miss.
Redundancy illustrates the point well. Under the previous arrangement, resilience meant a second physical line and perhaps a spare card in the switch. Now it means a secondary internet connection, battery backup for network equipment, and routing rules that divert calls to mobiles when a site becomes unreachable. Same objective, entirely different shopping list, and the second version is generally cheaper than the first.
Administration changes hands too. Enrolling a new joiner, changing a menu message or rerouting a department typically passes from a telecoms contractor to an administrator in your own office, which saves money and quietly creates a new responsibility. Name the person holding it before launch, rather than finding out afterwards that nobody does.
What you actually get
Core capability is fairly consistent between providers, because the category matured some years ago. Extensions and direct numbers for staff. Voicemail, often with transcription. Transfers, hold, forwarding and simple menus. Presence showing who is available. Mobile and desktop apps carrying the same identity as a desk phone.
Beyond that baseline, offerings separate. Some bundle video conferencing and team messaging, moving toward a full collaboration suite. Others add light call queueing. A few connect to business systems so a customer record opens alongside an incoming call, and roughly at that point this category starts brushing against customer-operations software.
Integration depth is worth probing during selection, since “integrates with your CRM” covers an enormous range. At the shallow end, a link opens a browser tab. Deeper implementations write call outcomes back automatically and pop a screen before anyone says hello. The gap between those two experiences is large, and both get described with identical marketing language.
Mobile behaviour rewards testing rather than reading about. How the app handles a call arriving while somebody is already on their personal line, whether it survives a phone restart, and how battery consumption looks over a working day are all things that separate products, and none of them appear on a feature grid.
Flexibility is the benefit people notice first. Staff answer the published number from home, from another emirate, or from an airport, with no forwarding arrangements and no second handset. For organisations running distributed or hybrid working, that alone tends to settle the business case.
Scaling behaves differently too, in both directions. Adding twenty people during a busy season takes an afternoon, and removing them afterwards takes less. Seasonal businesses and project-based teams find that considerably more useful than the marketing suggests, since the old model priced capacity for the peak and carried it all year.
Cost predictability comes second. A monthly per-user charge replaces capital expenditure, an annual support agreement and an unpredictable replacement cycle. Finance teams generally prefer that shape, even where the total across five years works out comparable.
Continuity of capability is the quieter third benefit. Because the provider upgrades the platform centrally, features arrive without a project, and the system does not gradually fall behind the way owned equipment does between refresh cycles. The trade is control over timing: upgrades happen when the vendor schedules them, occasionally changing an interface your staff had grown used to.
Reporting deserves noting too, though expectations need managing. Most systems in this category show you call volumes, durations and missed calls. That is genuinely more than an old switch offered, and considerably less than an operation with service-level commitments requires.
Where the limits sit
Three constraints deserve stating plainly, because vendor pages tend to underplay them.
Connectivity becomes a single point of failure in a way it never was with analogue lines. An old corded phone drew power from the exchange and worked during a local outage. The FCC notes that some internet-based services stop working during power cuts where the provider offers no backup, and that emergency services access can carry limitations compared with older lines. Neither rules out business use; both need a settled answer during design.
Call quality becomes your problem as well. Jitter, packet loss and constrained upstream bandwidth all surface as complaints about crackling or dropouts, and they originate inside your own network considerably more often than at the provider. Quality of service configuration is not optional at any meaningful scale.
Vendor lock-in warrants a mention as well, being rarely discussed on the pages selling these products. Your numbers, your configuration and your recordings all sit with a supplier, and how easily each can be extracted varies enormously. Ask about exit before signing, because the answer costs nothing to obtain then and a great deal later.
Support quality differs sharply between suppliers, and proves hard to assess before you need it. Ask how incidents are escalated, what response times are committed in writing, and whether anyone answers outside normal working hours. Voice failures are visible to customers within minutes, so this deserves closer attention than most software procurement gets.
Then there is the ceiling. Telephony of this kind handles conversations between named people extremely well. It handles queues of strangers far less well, a separate job entirely, and the closing section covers where that boundary falls.
Cloud, on-premise or hybrid?
For most organisations replacing an ageing switch, cloud delivery is now the default rather than the adventurous choice. That was not true a decade ago, and the balance has moved for reasons that are worth understanding rather than simply accepting.
Several factors moved it. Connectivity improved to the point where voice quality stopped being a live argument in most locations. Distributed working stopped being exceptional. And the economics of maintaining ageing equipment worsened as the people who understood it retired.
Availability is another, and increasingly decisive. Traditional line types are being withdrawn across several major markets, and the timetables are published. In the United Kingdom, the House of Commons Library reports that landline services move to a fully digital network by January 2027, with the legacy public switched telephone network withdrawn. The same briefing cites Ofcom figures showing British customers still on that network falling from 5.2 million in mid-2024 to 3.2 million a year later.
Estonia and the Netherlands have already completed their switch-offs, with several larger European and Asian markets winding theirs down on separate timetables. None of this binds a business in Dubai, Doha or Riyadh, nor does any of it appear here as an obligation. The useful signal is commercial: as legacy platforms retire across major markets, manufacturer attention, spare parts and support expertise follow the traffic. Anyone planning a ten-year estate therefore has grounds to move that owe nothing to any local deadline.
When on-premise still makes sense
Genuine cases remain, and pretending otherwise would be dishonest.
Sites with unreliable connectivity are the clearest. Where an internet link drops weekly, moving voice onto it is a poor trade regardless of what the brochure says. Certain regulated environments impose data location requirements that a specific cloud arrangement may not satisfy, and those belong with your legal team, not a vendor.
Large estates holding recently purchased equipment also warrant patience. Where a switch went in two years ago and works perfectly well, replacing it early is a write-off dressed as modernisation. A few specialised connections to building systems, older alarm panels or industrial equipment likewise have no modern equivalent yet, and finding that out mid-project is expensive.
Worth separating genuine constraints from institutional reluctance, though. “We have always run our own” describes a habit, and a preference for equipment somebody can point at describes a comfort. Both are understandable and neither belongs in a business case dressed as a constraint.
Hybrid arrangements bridge most of these. Keeping local survivability at one site while running everything else as a hosted service is common, and so is migrating department by department instead of all at once. Neither approach shows a failure of nerve; both reduce risk across a large estate.
One caution about hybrids, though. Running two systems means administering two systems, and the interim state has a habit of lasting years longer than the plan allowed. Put a date on the second phase and hold to it, or accept openly that you are running a permanent split and resource it accordingly.
Costs, and what a migration involves
Pricing is usually per user per month, which makes comparison look simpler than it proves to be. Published rates rarely represent an actual bill.
Four things move the total. The licence tier, since capability is generally bundled into levels and the cheapest one often lacks something you need. Number rental, charged separately per line and varying sharply by country. Usage, covering per-minute destinations and messaging beyond any bundle. And one-off items including porting fees, handsets, professional services and network upgrades.
TabaTalk publishes its own plans on a per-user monthly basis with a contact-sales tier for larger deployments, the standard shape across this market.
Build a twelve-month model rather than trusting headline rates. Licensing looks cheap beside capital expenditure at first and accumulates steadily thereafter, so an honest comparison runs across the full life of whatever equipment you are retiring. Include the support contract that ends, and the callouts that stop being booked, because leaving those out understates the case considerably.
Watch the tier boundaries carefully during evaluation. Category pricing tends to place the capability most businesses genuinely want one level above the advertised figure, so the figure in a comparison table is frequently not the figure you will pay. Price the tier you actually need, not the entry rate.
Number rental is the line most often forgotten entirely. It recurs monthly against every published line, varies sharply between countries, and accumulates quietly across a growing estate. Organisations adding lines opportunistically frequently hold several nobody has used since a campaign finished two years earlier, so review the inventory annually and release what no longer earns its keep.
The migration sequence
Six steps, in an order that reflects where projects actually stall:
- Inventory what you have. Every number, every extension, every device, and everything else quietly sharing those lines: alarms, lifts, card terminals, door entry. This step prevents most unpleasant surprises and costs the least.
- Check the licensing position. Establish who is licensed to carry voice in each territory you cover, as the next section explains.
- Choose the tier and confirm number availability. Per country and per category, in writing.
- Start porting early. Moving established numbers depends on the releasing operator’s cooperation and routinely sits on the critical path for the whole project.
- Design routing before go-live. Hours, menus, overflow, escalation and language handling.
- Pilot with one team, then widen. Measure call quality honestly during the pilot rather than asking whether it feels fine.
Skipping that inventory is easily the commonest error, and omitting the licensing check is the costliest. Both get missed for the same reason: they are unglamorous, and neither involves talking to a supplier about features.
How long should the whole thing take? Platform activation is usually quick, and TabaTalk publishes an under-24-hour figure for going live on its own product, which is broadly typical of the category. Porting runs on a separate clock set by the releasing operator and the country’s rules, and complex estates measure that part in months rather than days. Communicate the longer shape internally at the start, because a go-live date announced on the software timeline gets missed on the numbering one.
Plan the parallel-running period as well. Keeping old and new active together for a few weeks costs a little and prevents the situation where a missed forwarding rule sends customers to a line nobody monitors any more.
What changes in the Gulf
Everything above is technically identical wherever you operate. The regulatory position is not, and it determines what you can buy in the first place.
Voice service in the Emirates is regulated at the service level rather than only at the numbering level. TDRA treats voice and video calling over IP as a regulated activity, deliverable by licensed service providers or in collaboration with them, and licensees must block traffic from applications falling outside compliance. Providers absent from the published roster may request an exemption at the authority’s discretion.
That reframes the shortlisting exercise entirely. Before comparing feature grids, establish that each candidate can lawfully carry voice everywhere you operate. An international vendor with an impressive product and no local arrangement does not qualify, however attractive the pricing page looks.
Ask for specifics and get them in writing. Which licensee carries the traffic? Under what agreement? What happens to your service if that relationship ends? A supplier unable to answer plainly is describing an intention, not an arrangement.
Outbound calling adds a further condition. Cabinet Resolution No. 56 of 2024 requires marketing calls to use local phone numbers issued by telecommunications companies licensed in the State, registered under the calling company’s commercial licence, and prohibits dialling from lines neither registered to nor owned by that business. Free zone entities sit inside that regime, which regularly surprises companies holding a DIFC or ADGM licence.
Local presence conditions are worth checking early too. Several regulators attach requirements to geographic ranges, so obtaining a number with a particular city’s code may depend on having an address there. That constraint rarely appears in a vendor’s coverage map and routinely surfaces during provisioning instead.
Data location questions come up frequently in regulated sectors here, and sit with your own legal advisers rather than with a vendor’s assurance. Where recordings or customer records must stay inside the country, confirm how a given arrangement handles that before it becomes a contractual assumption.
Confirm what the answer means in practice, too. A supplier holding data regionally differs from one holding it nationally, and the difference occasionally matters to a regulator even when it looks immaterial on an architecture diagram.
Two operational points round this out. Multilingual service needs designing rather than assuming: running Arabic and English through one queue is ordinary practice regionally and barely mentioned in material written elsewhere, so decide early whether callers select a language, whether the number dialled implies one, or whether every person handles both. And working patterns complicate coverage, since the federal working week, your private-sector customers’ schedules and an offshore team’s roster may all differ. Ramadan compounds it further, moving contact volume later into the day and shortening working hours. Check real historical data before committing to any staffing model, and revisit it annually, since the calendar moves.
Frequently asked questions
What are the disadvantages of a cloud phone system?
Dependence on connectivity is the main one, since voice now shares a path with everything else on your network. Power cuts affect service where no backup exists, emergency calling may behave differently from an analogue line, and call quality follows network conditions, unlike a dedicated circuit. Monthly licensing also continues indefinitely, unlike a capital purchase that eventually stops costing money. Each is manageable with planning, and none argues against the category as a whole.
Is a cloud phone system cheaper than a landline?
Usually yes over a full replacement cycle, though not always in year one. Subscriptions remove the upfront purchase, the annual agreement and the callout charges, which together often exceed licensing across five to seven years. International calling is typically far cheaper. Set against that, per-user charges continue indefinitely and connectivity work may be required first. Model both options across the lifespan of the hardware being replaced and not on monthly figures alone.
Are there free cloud phone systems?
Free tiers exist, though they rarely suit a registered business. Limitations usually include a single user, restricted countries, no service commitment, and registration to a person, not to a company, which matters where numbering must sit under a trade licence. Call recording, reporting and support are commonly absent too. Treat free offerings as personal tools or short trials of a concept, not as infrastructure your customers depend on reaching you through.
Do I need special handsets, or will existing desk phones work?
It depends on their age, and many organisations need no handsets at all. Phones already speaking modern internet protocols generally work, while older analogue handsets need an adapter or replacing outright. Plenty of teams skip physical devices entirely, using laptop software and mobile apps instead, which removes the hardware line from the budget. Check what you own before assuming either outcome, since the answer changes the upfront cost materially and shapes the whole hardware budget.
Is a cloud phone system suitable for a very small business?
Yes, and small teams often benefit proportionally more than large ones. Two or three people gain a professional published number, a menu, voicemail and the ability to answer anywhere, without buying equipment or employing anyone to maintain it. Per-user pricing scales down cleanly, so the entry cost stays modest at a handful of seats. One caution applies at any size, though: confirm the provider may lawfully deliver voice locally before signing anything.
Where TabaTalk fits, and where to go next
Worth restating the scope point from the opening, since it shapes what this article could usefully tell you. TabaTalk sells cloud contact center technology, not a cloud phone system. The platform routes conversations, dials them, records them, analyses them and passes data onward. It holds no network, does not act as a carrier, and is not a collaboration suite, so internal meetings and group chat stay wherever your organisation already runs them.
The distinction matters when you are shopping. If your requirement is staff reaching each other and customers reaching a named person, telephony of the kind described above is what you need. If somebody in your business answers for how quickly a queue of customers gets handled, that is a different product with routing, queueing, recording and reporting built for volume, and the two frequently run alongside each other.
Where does that leave a reader partway between the two? Usually needing both, licensed differently: everyone on telephony, and only the customer-facing team on the heavier platform. Our explainer on how contact center software works covers that boundary, and omnichannel conversations, routing design and live reporting show what the upper layer adds. For teams whose agents work from several locations, the remote setup covers how published lines and physical location come apart.
Working through a migration off legacy equipment, or trying to establish where telephony ends and customer operations begin? Those questions resolve faster in conversation than against any feature grid. Speak to sales or book a demo, and bring your current numbers, your call volumes and the list of countries you operate in.
