Customer stories
Bolna Runs High-Volume Enterprise Voice Workflows on Vobiz
Overview
Business model: A B2B voice AI orchestration platform for enterprise customers, spanning e-commerce, edtech, BFSI, and health tech.
Key use cases: Collections, lead qualification, cart abandonment, customer support, KYC reminders, and appointment management.
Vobiz capabilities: TRAI 140/160 numbers, SIP trunks for outbound and inbound calling, and Voice APIs — programmable call control, live transfers, DTMF capture, recording, and post-call webhooks.
600k+
peak daily call volume
1400+
peak concurrent sessions
340+
DIDs provisioned
Partner
Bolna
Vobiz Products
Integration
Bidirectional WebSocket streaming
Geographies
Bolna is a YC-backed (F25) horizontal voice AI orchestration platform that helps enterprises run inbound and outbound voice workflows at scale. Vobiz provides the telephony layer to power these: delivering not only the numbers, routing, and media handling, but also high-concurrency infrastructure, and live routing across geographies.
The Challenge
Bolna's enterprise clients across e-commerce, edtech, BFSI, and health tech, use its platform to deploy AI agents that run outbound campaigns and handle inbound calls at scale. The use cases it supports present either infrastructural complexity, requiring voice and data flows across systems in real-time, or have demanding scale and speed requirements. This is where the telephony layer plays an important role.
Before onboarding with Vobiz, Bolna was already running its programs with two competing telephony players. As campaign volumes grew, two things led the team to look for a partner who is better aligned with their needs: the economics of high-concurrency dialing, and access to a wider range of number series that help run campaigns without being flagged as spam.
What Bolna was Looking For
Bolna evaluated Vobiz on three practical criteria: concurrency economics, number series access, and support.
Accessible concurrency economics - Bolna is not running occasional voice traffic. It runs concentrated daytime batch campaigns and needs the economics to hold as volume increases. Vobiz made that possible in a way the team found commercially better than the alternatives.
Breadth of number series - Standard geographic numbers are not enough for all campaigns. Bolna needed TRAI-mandated 140 and 160 series and local number provisioning as part of an outbound strategy that could support enterprise clients while reducing spam risk and improving answer rates. Vobiz offers deep number series pools across 190+ countries for outbound calling.
Responsive support and development - Unresolved telephony issues become visible immediately in campaign performance. Vobiz offers reliable and responsive support quality even as volumes scale, backed by an agile product development roadmap that is constantly revamping the traditional telephony layer so that it is more suitable for a voice-first company.
What Bolna Built on Vobiz
With Vobiz, our teams get programmatic access to telephony capabilities that used to require months of vendor coordination and specialized telecom know-how. Clean IVR-AI-Agent transfers, post-call delivery signals, and live cross-region routing are now API primitives for us, not bespoke infrastructure projects.
“-Maitreya Wagh, Bolna
Three Vobiz capabilities are proving essential to Bolna's most complex production flows.
IVR + AI + Human agent flows on one call leg
In multi-leg inbound flows, the most common failure is context loss at handoff: when a call moves from IVR to an AI agent to a human agent, each leg starts cold unless the platform explicitly preserves and passes state.
Vobiz enables this for one of Bolna's enterprise clients, who runs a combined flow across IVR, AI agents and human agents. The customer's call is picked up by a Vobiz IVR, verified, and then routed to an AI agent, a human agent, or a ticket on the client's platform based on the intent detected. Bolna programs this routing via Vobiz's call control API, with transfers carrying the full conversation context forward across every leg. Each agent picks up seamlessly from the last one, and customers don't have to repeat themselves.
Clean call endings via post-call event signals
In AI-driven calls, there is a meaningful gap—typically 200 to 500 milliseconds—between when the AI finishes generating a response and when the final audio packet finishes playing for the customer.
On platforms that do not signal delivery, the backend has no reliable way to know when the final audio has finished playing. So, it either waits for a timeout or leaves the call open indefinitely. The customer has finished listening, but the line stays active, creating seconds of blank silence before the call eventually drops.
Bolna uses Vobiz's post-call event streaming to end calls smoothly just as the final message has fully played out. Calls end naturally, without awkward silences.
Live cross-region routing without dropped calls
Caller ID origin has a direct impact on answer rates in certain use cases where customers expect calls from international numbers, or in outbound campaigns targeting users in other countries. Also, provisioning international numbers is only part of the challenge. When a live call needs to transfer from an international agent to an Indian counterpart mid-conversation, the server routing has to switch geographies in real time, without dropping the call or requiring reconnection.
For a remittance payments client, Bolna provisions US numbers through Vobiz for outbound origination. When a call needs live transfers between US and Indian contacts, Vobiz handles the server switch seamlessly, and in real time.
About Bolna
Founded in 2024, Bolna is building a self-serve platform that allows enterprises to design, deploy, and monitor voice AI agents, without long implementation cycles or specialized AI expertise. The platform supports more than 10 Indian languages and is designed for real-world telephony conditions, including noisy environments and regional accents. bolna.ai
About Vobiz
Vobiz is a programmable voice infrastructure platform built for voice-first products. It provides the telephony layer — number provisioning, bidirectional WebSocket audio streaming, programmable call control, TRAI compliance infrastructure, and post-call event streaming — that AI orchestration platforms and enterprise product teams rely on to run production-grade voice operations at scale.