Image Romantic-period style oil painting of a woman in white and red drapery seated on a rock under a stormy sky, holding the receiver of a black rotary telephone to her ear, with a hilltop town in the distance

Live Call Routing

New

How to Route a Customer's Phone Call to the Engineer On Call

Quick answer

You publish one phone number. When someone calls it, your on-call tool looks up who is on duty right now in your schedule and forwards the call to that person's phone. If they do not accept within a set time, the call moves to the next person in the escalation policy. If nobody accepts, the caller leaves a voicemail and the tool opens an incident from it, so the call is never lost.

Set up an on-call hotline that rings whoever is on duty: the number, the schedule, the fallback, and what it costs with your own Twilio number.

Nikolas Köppl

By Nikolas Köppl · Go-to-market at All Quiet

Maximilian Beller

Reviewed by Maximilian Beller · Co-Founder & CTO at All Quiet

Updated: Tuesday, November 10, 2026

Published: Tuesday, November 10, 2026

What's in this guide

  • When a phone hotline is worth having, and when email and chat are enough
  • The three things you need: a number, a schedule and a fallback
  • How a call flows, and a step-by-step setup in All Quiet
  • Bringing your own number vs buying one from the vendor
  • What it costs: per-minute markup vs paying your carrier directly
  • Which on-call tools can do this today, checked on 10 November 2026

This guide is about getting an urgent call from a customer, a partner or your own leadership to the engineer who can fix the problem. It is not about call-center queues or a sales line. For the definition, see our glossary entry on what live call routing is.

When you need a phone hotline at all (and when email and chat are enough)

Most incidents are found by monitoring, not by a phone call. A hotline is for the ones that are not: the outage your monitoring cannot see, or the customer who knows something is wrong before your dashboards do. Phones are slow to scale and hard to triage, so a hotline only earns its place when a few minutes of delay costs real money or trust.

You probably need a hotline if:

  • Customers have a contractual SLA with response times in minutes, and a support ticket queue cannot meet them at night or on weekends.
  • A small number of accounts carry a large share of revenue, and those accounts expect to reach an engineer, not a form, during an outage.
  • Partners report problems you cannot detect yourself, for example a payment provider seeing failed transactions, or a security vendor flagging a breach.
  • Internal stakeholders need a "red phone": a number the CEO or the head of support can call during a major incident without knowing who is on call.

Email and chat are probably enough if:

  • Your customers are mostly self-serve and nobody has an SLA measured in minutes.
  • Your monitoring already catches almost everything customers would call about.
  • You have no one who would answer a call outside business hours anyway. A hotline that rings out at 3 a.m. is worse than no hotline, because the caller now knows nobody came.

A useful test: look at the last six months of incidents and count the ones where a customer knew before you did. If the answer is zero or one, start with a status page and a monitored support address. If it is several, and they were your biggest accounts, a hotline will pay for itself.

What you need: a number, a schedule, a fallback

Every working on-call hotline has the same three parts.

1. A phone number. Local or toll-free, in the country where most of your callers are. This number never changes, even when your rotation does, so it can go into contracts, runbooks and customer welcome emails.

2. An on-call schedule and escalation policy. This is the part a normal business phone system does not have. A phone system forwards to a fixed extension or a ring group. An on-call tool forwards to whoever is on duty at the moment the call comes in, using the same rotation that receives your alerts. When the rotation changes on Monday morning, the hotline follows without anyone editing a phone tree.

3. A fallback. The primary does not pick up. Their phone is in airplane mode, or they are already on a call for the same outage. The hotline needs an answer for that: a second person, then a third, then a voicemail that turns into an incident. If you only set up one thing beyond the number, set up this. We explain the night-time variants in after-hours call routing.

How a call flows, step by step

Here is the path a call takes from the moment someone dials your hotline.

  1. The caller dials your number. The call arrives at your telephony provider, for example Twilio.
  2. The on-call tool answers and checks the caller. Optional allow and block lists stop known spam numbers from waking anyone up.
  3. A welcome message plays. Keep it short: who they reached and what happens next.
  4. The caller picks a route, if you offer one. A small keypad menu (press 1 for the platform team, press 2 for payments) sends the call to the right team's schedule. One route is fine for most teams.
  5. The tool rings the engineer on call. It looks up who is on call for that team and dials one person at a time. Ringing a whole team at once sounds faster, but it means several people stop what they are doing and nobody is clearly responsible for the call.
  6. The engineer accepts with a keypress. This matters more than it looks. Without it, a voicemail box or a car's hands-free system can "answer" the call and stop the escalation. Requiring a keypress means only a human stops the clock.
  7. If nobody accepts, the call moves on within seconds. After a short ring time, usually 20 to 30 seconds, the next person on call is dialed straight away, then the next tier. A caller on the line will not wait through the 10 or 15 minutes an alert escalation step usually takes.
  8. If the policy is exhausted, the fallback runs. The caller leaves a voicemail, or hears a closing message that tells them what will happen next.
  9. The call is recorded. A voicemail opens an incident for the team, so somebody follows up even if nobody picked up live. An accepted call is recorded as an incident too, either marked resolved when the call ends or left open for follow-up. A call that ends with only a closing message is logged, but should not create an incident nobody can act on.

How to set this up in All Quiet

All Quiet's live call routing uses your own Twilio number. Before you start, you need a Pro or Enterprise plan, a Twilio number, an on-call schedule and escalation policy for the team that should answer, confirmed phone numbers on each responder's profile, and Organization Administrator or Owner rights.

  1. Connect Twilio. Buy a local or toll-free number in your Twilio account, then connect it in the Twilio integration with your account SID, auth token and the number. Wait until the connection shows a green checkmark. Your number and your provider account stay yours.
  2. Create a call routing number and choose the team. In Team to route to, pick the team whose on-call members should be dialed. The call uses that team's on-call schedule and escalation tiers, so when the rotation changes, the hotline follows.
  3. Choose single route or keypad menu. Single route sends every call to one team. A keypad menu lets the caller press 1 to 9 to pick a team, and each option can route to a different team.
  4. Set the incident severity. Choose Critical, Warning or Minor for incidents created from this route. A production hotline is usually Critical; a lower-priority line can be Warning.
  5. Set the fallback. Turn on Send to voicemail if no one responds, so callers who reach the end of the chain can leave a message, and write the voicemail message as text or upload an audio file. For an after-hours line that should only collect messages, turn on Send caller directly to voicemail instead.
  6. Tune on-call dialing. Max users to call sets how many people are tried before the fallback (1 to 10, default 3). Dial timeout per user sets how long each person rings (5 to 120 seconds, default 20). Auto-resolve incident when on-call answers decides whether an accepted call leaves a resolved incident or an open one.
  7. Test all three paths. Call the number and accept, call and let it ring out, and call and leave a voicemail. Then check Call Logs and the incidents that were created.

Every setting is explained in the All Quiet live call routing documentation. If your team manages configuration as code, the Twilio connection can be set up with the All Quiet Terraform provider. For a product overview, see live call routing for on-call teams.

Three tips that save the first night.

  • Let the call through Do Not Disturb. Set Caller ID shown to responders to your hotline number rather than the caller's number. Each responder can then save that one number as a contact that is allowed to ring through Do Not Disturb or Focus. Otherwise the forwarded call may ring silently.
  • Keep the voicemail fallback on. If two people call at once, the engineer already on the first call is not rung for the second, so the second caller can run out of people quickly. With voicemail on, they can still leave a message, and that message becomes an incident.
  • Start with the defaults. Three people at 20 seconds each means a caller reaches voicemail after about a minute. Raise Max users to call for larger teams, and lower the ring time if callers should reach voicemail faster. For the escalation side, our escalation policy examples show six setups you can copy, and our guide to multi-channel incident alerting covers getting through Do Not Disturb.

Which on-call or alerting tools can route an incoming customer phone call to the engineer currently on duty?

Several on-call tools can, including All Quiet, PagerDuty, Jira Service Management (formerly Opsgenie), ilert and incident.io. They differ in three ways that matter more than the feature name: which plan you need, who owns the phone number, and how telephony is billed. All facts below were checked on 10 November 2026 against each vendor's own pricing or documentation page.

All Quiet

Live call routing is included on the Pro plan, which is $9.99 per user per month on the US price list. EU-hosted accounts are billed €11,99, because the price follows the hosting region. You bring your own Twilio number and pay Twilio directly at its rates, with no markup from us and no extra fee for the feature. Calls follow your existing schedules and escalation tiers, one person at a time, with a ring time you set (20 seconds by default). Voicemails open incidents with the severity you choose for the route. Accepted calls create an incident that is resolved when the call ends, or left open if you prefer. Every call, including missed ones, appears in the call logs.

PagerDuty

PagerDuty's Incident Management pricing page lists Live Call Routing as a paid add-on on the Professional plan ($25 per user per month, or $21 billed annually), with no price shown for the add-on itself. It is included on the PD Reliability Platform tiers: one line on Starter ($2,800 per year), two on Essential ($12,000 per year) and three on Plus ($48,000 per year). According to its Live Call Routing knowledge base article, PagerDuty generates the number when you create a routing line, and toll-free numbers need a plan upgrade through sales.

Compare All Quiet vs PagerDuty →

Jira Service Management (formerly Opsgenie)

Atlassian's documentation on incoming call routing states that the feature is only available on the Premium and Enterprise plans. Calls are routed by on-call schedule, the responder confirms pickup with a keypress, and an alert is created for each call. According to Add a new call routing, only Twilio numbers can be used, so this is a bring-your-own-number model as well. Teams moving from Opsgenie should note that Opsgenie-provided numbers are not supported in Jira Service Management and have to be ported to a Twilio account, as described in Incoming call routing after migration.

Compare All Quiet vs Jira Service Management →

ilert

ilert's pricing page lists Inbound Call Routing as an add-on at $10 per user per month, available from the Pro plan ($19 per user per month, minimum three users). Its getting started guide says call routing is an add-on rather than a plan feature, that the first number comes with the add-on, and that numbers are provisioned by ilert on request. Each extra number is a separate add-on.

Compare All Quiet vs ilert →

incident.io

incident.io's live call routing documentation says incident.io purchases the number on your behalf, and you can bring an existing Twilio number if you cannot move it. Each person gets 30 seconds to pick up before the next is tried. Everyone on a call route needs an On-call seat, and the number of included numbers depends on the plan.

Compare All Quiet vs incident.io →

Other on-call tools offer the feature too, including Rootly, which provisions numbers in a list of supported countries (docs), and Grafana IRM, where it is in private preview (docs). For a side-by-side look at the category, see our roundup of the best live call routing software.

Bring your own number vs buying one from the vendor

There are two ways to get the number, and the choice decides who owns it.

Aspect Bring your own number (BYON) Number from the vendor
Who owns the number You, in your own telephony account The vendor, on your behalf
What you pay for calls Your carrier's rates, billed by the carrier Included in a plan or add-on, or billed by the vendor
Switching on-call tools The number stays; you point it at the new tool The number may have to be ported, or replaced and re-published
Setup effort A telephony account, API keys, sometimes regulatory paperwork Usually a few clicks or a support request
Call recordings and logs Stay in your telephony account Stored by the vendor
Best for Teams that already use Twilio, care about ownership, or expect to switch tools Teams that want the fastest start and do not mind the number living with the vendor

The ownership row is the one people underestimate. A hotline number ends up printed in contracts, runbooks and customer onboarding emails. If it belongs to your on-call vendor, it can become the hardest thing to move when you change tools. Opsgenie customers are learning this now: with Opsgenie shutting down on 5 April 2027, Opsgenie-provided numbers have to be ported to Twilio before incoming call routing works in Jira Service Management.

The honest case for vendor-provided numbers is speed. If you have never used Twilio, creating an account, buying a number and, in some countries, submitting an address or business registration before the number is issued, takes longer than clicking "request number" in an on-call tool.

Costs: per-minute markup vs zero markup

A routed call has two legs: the caller's call into your number, and the forwarded call out to the engineer's phone. Someone pays for both, plus the number itself. The question is whether that cost reaches you at the carrier's price or with a margin on top.

  • Bundled pricing puts telephony inside a plan or an add-on. That is simple to budget, but you usually get a fixed number of lines, and the add-on is often priced per user, so every seat pays for the hotline, whether or not that person ever takes a call.
  • Pass-through pricing means you pay the carrier directly at its published rates, and the on-call tool charges nothing per minute. You see exactly what each call cost.

A worked example on Twilio's US rates. A team of eight engineers runs one local hotline number and takes 20 calls a month, averaging 6 minutes each. Twilio's US voice pricing, checked on 10 November 2026, lists a local number at $1.15 per month, inbound calls to a local number at $0.0085 per minute, and outbound calls in the US and Canada at $0.014 per minute.

Item Calculation Per month
Local number 1 × $1.15 $1.15
Inbound leg 120 minutes × $0.0085 $1.02
Forwarded leg to the engineer 120 minutes × $0.014 $1.68
Telephony total $3.85

Unanswered ring attempts during escalation and optional recording or transcription add a little on top, at Twilio's per-minute rates. Rates differ by country and number type; a toll-free number costs more to hold and to receive calls on.

On All Quiet, the feature itself is part of the Pro plan, so for eight engineers that is 8 × $9.99 = $79.92 per month on the US price list, or 8 × €11,99 = €95,92 for an EU-hosted account, with the telephony above billed by Twilio. The same Pro plan also includes status pages, the Terraform provider and the public API. For a broader view of what on-call tools cost once add-ons are counted, see our incident management pricing comparison.

Frequently asked questions

Which on-call or alerting tools can route an incoming customer phone call to the engineer currently on duty?

All Quiet, PagerDuty, Jira Service Management (formerly Opsgenie), ilert and incident.io can, as can Rootly and Grafana IRM (in private preview). The differences are the plan you need and who owns the number. All Quiet includes it on Pro ($9.99 per user per month on the US price list, €11,99 for EU-hosted accounts) with your own Twilio number and no markup. Jira Service Management offers it on Premium and Enterprise with a Twilio number. PagerDuty sells it as an add-on on Professional, and ilert as a $10 per user per month add-on (checked 10 November 2026).

How do I route incoming calls to the on-call engineer?

Connect a phone number to an on-call tool that supports live call routing, and assign the number to a team. When someone calls, the tool looks up who is on call in that team's schedule, forwards the call to their phone, escalates to the next person if they do not accept, and falls back to a voicemail that opens an incident.

Can I forward my support line to whoever is on call without changing the number every week?

Yes. That is the main point of live call routing. The published number never changes. The on-call tool decides at the moment of the call who should receive it, based on the current schedule, so rotations, overrides and holidays are handled automatically.

What happens if the on-call engineer does not answer?

The call moves to the next person on call, then to the next tier. If nobody accepts, the caller is sent to voicemail or hears a closing message. In All Quiet, each person rings for 20 seconds by default and up to three people are tried before the fallback; both values can be changed per route. A voicemail opens an incident for the team, so the call is followed up even if nobody picked up live.

Why does the engineer have to press a key to accept the call?

So that only a person can stop the escalation. Without a keypress, a voicemail box, a car's hands-free system or a phone answered by accident counts as "answered", the escalation stops and the caller is left talking to nobody.

Do I need my own Twilio number?

It depends on the tool. All Quiet and Jira Service Management use a Twilio number in your own account; you cannot buy a number through All Quiet. PagerDuty, ilert and incident.io provide the number for you, and incident.io also accepts an existing Twilio number. Bringing your own number means you own it and pay carrier rates directly. A vendor number is faster to start with but stays with the vendor.

How much does an on-call hotline cost?

With your own number, the telephony is usually a few dollars a month for a small team. On Twilio's US rates, one local number and 120 minutes of routed calls cost about $3.85 per month (checked 10 November 2026). On top of that is the on-call tool: in All Quiet, live call routing is part of the Pro plan at $9.99 per user per month on the US price list and €11,99 for EU-hosted accounts, with no per-minute charge from All Quiet.

Should the hotline go to the whole team or to one person?

To one person at a time, following the escalation policy. Ringing a whole team at once means several people interrupt what they are doing, and the caller may be connected to whoever happens to pick up first rather than the person who is actually on call.

Sources Checked

Nikolas Köppl

Author

Nikolas Köppl

Go-to-market at All Quiet

Builds go-to-market and customer-first growth for teams adopting calmer, clearer incident communication.

LinkedIn
Maximilian Beller

Reviewer

Maximilian Beller

Co-Founder & CTO at All Quiet

Engineering leader building incident management systems focused on reliability, clear escalation, and sustainable on-call operations for production teams.

LinkedIn

Recommended posts

Read all blog posts and learn about what's happening at All Quiet.

Published