Health checks
Every trunk, SIP edge and media pod is probed continuously, so a line that is quietly failing is found by a check rather than by a customer.
Cette page est disponible en anglais pour le moment. Voir la page en anglais
Telephony & voice infrastructure
Voice AI telephony for production calling workflows
A voice agent is only as good as the call it rides on. Dring connects your existing or provisioned numbers, SIP trunks, virtual PBX, recording and routing to the agent. Or we peer with the switch you already have, so the first workflow can fit the lines and controls your team already operates.
The telephony layer
Carriers in front, a media cluster behind.
Traffic from carriers and your own switch is filtered by line protection, terminated on a replicated SIP edge, and carried by a media server cluster that runs as replicated workloads and auto-scales with concurrent calls. Health checks run on every trunk, SIP edge and media pod, alerting the on-call engineer when one degrades, with active and standby paths scoped to the deployment. The voice pipeline sits behind all of it.
Read left to right. A call arrives from a carrier or your PBX, passes line protection, is terminated on the SIP edge and anchored on a media pod, where it is recorded and where DTMF and IVR input are handled. The primary path carries traffic and the configured standby path can take over if the deployment supports it. Only then does the voice pipeline speak.
Operations
Alerts
Degradation pages the on-call engineer before customers notice. A trunk losing calls, a SIP edge refusing registrations or a media pod struggling all raise the same alert path.
Auto-scale
Media capacity follows concurrent calls, up and down. A campaign that doubles its dialing does not need a hardware order, and quiet hours do not pay for peak.
Numbers and lines
Local numbers where your customers are. Or the numbers you already own.
Two ways in, both production-tested: connect a number already provisioned for your business, or peer with the telephony you already run. Dring manages the routing and setup around the workflow.
We connect and run the line
Your carrier or telecom partner provisions the number. Dring connects it, configures queues, routing and business-hour rules, and manages the setup for your workflow. Recording and carrier redundancy are configured per deployment; number porting is available when you move an existing line.
We peer with your switch
Your PBX or contact centre platform stays the system of record. Dring connects over SIP as another destination: route a queue, a menu option or an after-hours schedule to the agent, keep everything else as it is. Calls the agent cannot finish come back to your queues warm, with context attached.
IVR
IVR that understands sentences.
Callers say what they need instead of picking a number from a menu. The agent listens to the sentence, works out the intent, and routes the call or handles it directly. A classic "press 2" tree stays available on the lines that still need one, alongside the natural-language flow.
- Natural-language routing: the caller speaks, the agent understands and routes
- Optional classic DTMF tree on the same line, where it is still needed
- Business-hours and holiday routing, calendar-aware
- Queue callbacks instead of holding on the line
- Language detection at the greeting, before the caller chooses one
- Hand-off to a human queue with full context attached
AgentThanks for calling. What can I help you with today?
CallerI have a question about an invoice I received.
Intent detectedInvoice question
Routed toFinance agent
Transfer to humanIf the caller asks, with context
Virtual PBX
Virtual PBX features
What a contact-centre PBX does, run as configuration rather than a box in a closet.
Queues and ring groups
Calls held and distributed across a group of agents or lines, in the order and priority you configure.
Business hours and holidays
Routing follows your calendar: opening hours, weekly closures and holiday schedules, per line.
Voicemail detection and drop
Answering machines detected on outbound calls, with a message dropped by policy instead of talking to a machine.
Call recording with retention
Recording can be configured for the workflow, with access, retention and deletion rules set during deployment.
Conference and warm transfer
Bring a third party onto the call, or hand off to a human with the context already shared.
Extension dialling
Reach your own team by extension, from the agent or from a caller routed internally.
Numbers and messaging
Numbers and messaging
The line you already have and the channels around it, connected and managed in one place.
Local and toll-free numbers
Your carrier or telecom partner can provision local and toll-free options; Dring connects and manages the configured line for the workflow.
Number porting
Move an existing number onto Dring telephony without changing what your customers dial.
Caller ID management
Set the outbound caller ID per campaign, so customers see a number they recognise and can call back.
SMS gateway
Links and confirmations sent by SMS during or after the call, in the caller's language.
WhatsApp Business
Dring operates as a Meta WhatsApp tech provider, with template approvals handled for you.
Email notifications
Summaries and confirmations sent by email to your team or the customer.
Line analytics
Line analytics
Every line measured on its own, so one bad number or one bad hour does not hide behind a good average. Full detail on the analytics page.
Answer rate per line
Share of inbound calls answered, tracked per number.
Abandonment before answer
Calls the caller ends before the agent picks up, a sign of queue or wait-time trouble.
Hold and queue times
How long callers wait in a queue or on hold before being answered.
Out-of-hours share
Share of call volume landing outside business hours, useful for staffing and routing decisions.
Per-number quality signals
Audio quality and latency signals tracked per number, so a degrading line is caught early.
Outbound discipline
Outbound that respects the rules of every country.
Outbound campaigns are governed by configuration, not by hope. Every rule is enforced at the edge before a number is dialed.
Calling windows per country
Permitted hours and days follow the destination's rules, your consent model and the customer's time zone. The configured window is checked before dialing.
Re-dial and retry rules
Retry count, spacing and daily caps per campaign. Voicemail detected and handled by policy. Unanswered numbers are re-dialed until there is an outcome or the limit is reached, never beyond.
Do-not-call and consent
DND lists honoured before dialing, consent recorded in the transcript, opt-out kept per number. Caller ID is set per campaign so customers see a number they can call back.
Caller identity
Inbound callers are matched to your CRM by number before the agent speaks. Unknown numbers get a verification flow you define; known customers skip it.
Recording and retention
Recording and transcription are configured per workflow. Retention, access and deletion follow the notice and operational rules agreed for the deployment.
Capacity on demand
Concurrent calls scale with your campaign, not with a hardware order. Peak-hour inbound and a same-day outbound campaign run side by side on separate lines.
PBX and SIP security
Line protection
The line itself is a target before an agent ever answers. Protection is configurable per line, as agreed in your contract.
SIP brute-force and scanner blocking
Repeated failed registrations or malformed requests trigger an automatic ban on the source, in the style of fail2ban.
Spam and robocall flood protection
Rate limits per source and burst detection quarantine suspicious traffic automatically, so real callers keep getting through.
Caller ID validation
Spoofing checks run on every inbound call. Unknown or unverified numbers are routed into the verification flow you define.
Toll-fraud controls
Destination allow-lists and spend caps per line and per campaign, with alerts on anomalous call patterns.
Encrypted signalling and media
TLS on SIP signalling, and SRTP for media where the carrier supports it.
Incident response
Defined severity levels, a named owner, and customer communication within the first hour of a confirmed incident.
Beyond voice
WhatsApp, SMS and email on the same trunk of truth.
Dring is a Meta WhatsApp tech provider, so a call can continue on WhatsApp for a photo or a document and return to voice for the confirmation. SMS carries links and confirmations; email carries summaries. Same agent, same memory, same case.
- WhatsApp Business messaging with template approvals handled for you
- SMS links, confirmations and reminders in the caller's language
- Email summaries and notifications to your team or the customer
- One conversation record across all of them
VoiceCaller asks to return an item. Eligibility checked.
Agent · WhatsAppYou can send the photos in this chat. I will review them right away.
WhatsApp2 photos received · return approved
SMSShipping label link sent
EmailSummary to the support inbox
Operations
Watched around the clock, by people who answer.
Infrastructure is a service, not a shipment. Monitoring, alerts, staged rollouts and a named client operations contact come with every line. Compare the wider picture in the voice AI operations snapshot, review the Agent Factory operating loop and use the 62-language customer service guide for language planning.
300Kabout customer conversations/month across inbound+outbound
10K+agent conversation minutes/day
72%about resolution across current production workloads
62language technical coverage, with 10 launch-priority languages
- Monitor
Live line health
Answer rates, audio quality signals and latency per line, watched continuously. Alerts go to Dring first, then to you if action is needed.
- Roll out
Controlled stages
Every change to an agent or a route goes live in controlled stages, with the previous version available for rollback.
- Respond
Incident playbook
Defined severity levels, a named owner, customer communication within the first hour of a confirmed incident.
- Review
Monthly report
Volumes, answer rates, outcomes and quality trends per line, with the actions we took and the ones we propose.
FAQ
Telephony questions.
Can we keep our existing numbers and PBX?+
Yes. Dring connects to your switch over SIP as a destination. You route what you want to the agent and keep the rest unchanged. Porting numbers to Dring is optional.
Can we keep our existing IVR?+
Yes. Dring can sit behind one option of your current IVR, taking only the calls you route to it, or replace the menu entirely with a natural-language flow that still falls back to a classic tree where you need one.
How do we get a local number for our market?+
Local and toll-free availability depends on the destination, carrier and use case. Your carrier or telecom partner provisions the number; Dring connects it and manages the routing and setup. Share the market and workflow through the scoping form so we can confirm the right path, porting and caller-ID options.
What happens if a carrier or edge path has an outage?+
Traffic can move to the configured standby path automatically, with carrier redundancy scoped per market. Calls in progress on a failed leg may be affected; new calls continue on the healthy path where the deployment supports it.
How are outbound calling rules enforced?+
At the edge, before dialing: calling windows per country and time zone, retry caps, do-not-call lists and caller ID policy. The agent cannot place a call the configuration forbids.
Where are calls processed and stored?+
Processing location, transfers, recordings, transcripts and retention are reviewed for the deployment and documented in the agreed controls. Details are on the security page.
How do you protect the lines from SIP attacks and spam floods?+
Automatic blocking on repeated failed registrations and malformed requests, rate limits and burst detection against spam and robocall floods, caller ID validation, toll-fraud controls and encrypted signalling and media, configurable per line. See line protection.
Bring us your provisioned line.
Tell us about numbers, routing, integrations and the go-live plan. A consented form submission can be followed by an AI qualification callback in about two minutes, with timing confirmed for the workflow.