On-premises phone systems: equipment plus subscription
A traditional phone system relies on equipment installed on the premises, plus handsets, installation and maintenance. Spending is therefore split between an upfront investment and recurring costs: the operator subscription, a maintenance contract, call-outs when something breaks.
This model still makes sense for some organisations committed to on-site infrastructure, but it leaves the hardware risk with the business: ageing equipment, upgrades to fund, dependence on an installer. It is the model with the heaviest share of fixed costs.
Virtual phone systems: a price per user
A virtual or cloud phone system does away with equipment: everything runs over the internet, and each team member takes calls on their computer or mobile. Billing is typically a monthly subscription per user, including the number, call routing, queues and basic answering features.
The bill therefore grows with the size of the team and the tier chosen: some features (recording, analytics, integrations) are reserved for higher tiers. An important point: a virtual phone system routes calls but doesn’t answer them. If nobody is available, the call ends up in voicemail.
Human outsourcing: by time or per call
An outsourced answering service has your calls picked up by virtual receptionists. Two billing models dominate: a package including a volume of calls or time, with overage billed on top, or a charge per call handled. Extended hours (evenings, weekends, public holidays) are sold as options and push the price up significantly.
Beyond the headline rate, check what counts as a handled call, how very short calls or wrong numbers are treated, and the cost of setting up the brief. The price of this model tracks your volume mechanically: the more calls you receive, the higher the bill.
AI agents: subscription plus minutes
An AI agent that answers the phone is generally billed in two parts: a monthly subscription for the service (the number, the configuration, the call reports) and usage based on minutes of conversation. Some plans include a volume of minutes in the subscription; others bill usage separately.
What sets this model apart economically is that round-the-clock coverage costs no more: answering at night or on a Sunday requires no extra resources, unlike a human answering service. The cost follows actual conversation, not availability.
What drives the bill, whatever the model
Four factors weigh on every model: call volume, the hours of coverage required, the complexity of the handling requested (simple screening, detailed message-taking, appointment booking, qualification) and integrations with your tools (calendar, CRM). Before comparing prices, compare scopes: a cheaper offer that covers neither your hours nor your use cases will cost more in lost calls.
For Alyas pricing (subscription, included minutes and options), the up-to-date details are on our pricing page, and the 10-day free trial lets you estimate your real usage before committing.
Frequently asked questions
Which model is the most cost-effective for a small business?
It depends on what you’re missing. If your calls are already answered and you only need routing, a per-user virtual phone system is enough. If the problem is that calls aren’t being answered (people away, out on jobs, outside hours), you need a solution that actually answers: a human answering service or an AI agent, to be compared on your volume.
Are there hidden costs to watch out for?
The most common are set-up fees, overage charges, extended hours billed as options, advanced features reserved for higher tiers and minimum contract terms. Always ask what happens when your volume exceeds the package.
Why do providers so rarely publish their prices?
Because many offers are quote-based, depending on volume and scope. That is legitimate for highly customised services, but a public price list lets you compare with confidence: favour providers that are transparent about how they bill.